And don't forget that on every N step you have to _load_ more puddle. Also remember that it's not just about the width and depth, but also the _variety._ Color, flavour, atmosphere, etc.
I respect the hell out of this man for saying he doesn't play games he really wants to enjoy on stream. It's like people forget that streaming is a job for them.
Im mixed with this statement. On one hand, you got to play the game you love however you like. On the other, you practically admit that you dont enjoy the game that you actually stream.
@@RhazOfRheoshe didn't say that. Your assuming that from his words. I think he enjoys games in general then they're are games personal to him he wants that time to himself not sharing it with the stream.
Isn’t him saying this in relation to the fact that BG3 is a long game, he wants to finish it, and if he streams it he plays at half efficiency (due to like chat interaction and such)? It’s not that he can’t enjoy it, it’s just the game will take longer.
@@SeriouslyNotMichaelJordan he started playing Dragon Age Origins on stream a while back, but decided to finish it off stream so he could really get into it and enjoy the story without worrying about interacting with chat.
It's not really to do with it being a job, it's that it's harder to be truly immersed while you're streaming. This works like this even if you're streaming it to a single friend.
Starfield's ultimate problem is that it's not just a lot of nothing, but the fact there is stagnancy and corporate safeness. Like there are essential NPCs in places despite the entire game being about the multiverse, there aren't any mechs despite there being independent space, you can't edit the interior of your ship despite ships bing the main advertisement of the game.
Starfield's ultimate problem is that it's a Bethesda title that they "worked" on. Peel back the new coats of paint and wallpaper and it's literally a Skyrim clone with less content. The problem is the whole video game industry, mainly AAA stuff, has been going fully corporate for years now, and we're pretty much at the end game for that. I wouldn't expect anything traditionally good from them anymore.
I saw a great example of a random quest: Go find our missing scientist! He's 400 m away and you have to walk there. Nothing happens or gets found on the way. Arrive at the cave; dude is just standing there not in danger. No escort back; he just walks back on his own. Pass Go, collect $200. Skyrim's random quests pissed me off because you get the same pittance for killing a Dragon or a Bandit Chief at level 1 as at level 50, but at least there was some excitement, found a new shout word, found a new cave on the way, got some good treasure, whatever. Starfield seems even lazier than Skyrim.
Im feeling like this in almost every Bethesda game. They are only building a library of asset blocks and reuse them over and over again. Every house, every cave, every dungeon, sewer, cabin, subway looks and feels the same from the inside. And so do the 5 outposts and caves you find in starfield over an dover again.
In Starfield they didn't even bother moving the furniture. Literally. You can find the same notes and named corpses in copypasted facilities on different planets.
@@leinadreign3510 At least in most of their previous games they touched them up by hand or often had at least something interesting in it/going on or a quest related to it. Starfield is way worse because of the procedural generation you don't even have that much and random exploration feels a lot emptier because if you land somwhere random there isn't even the chance to stumble upon an interesting quest, item or location because the only hand made stuff is in the pre marked locations and quest locations. At least so far I haven't found anything in a random location I couldn't have found while doing the quests.
Hey , over at is having trouble with some and they've asked us to deal with it. Go do that, then come back to me for your reward. reward = levelled variant of some semi-random generic weapon
There's another story about Monty Python's The Holy Grail that I feel fits in well here: While filming one of the scenes in the swamp, production was slowed by a faulty fog machine. After a long while, a frustrated John Cleese yelled to the crew, "Is the fog funny?" The moral: if an element to your work of art doesn't exist to serve your ultimate point, then why worry about keeping it at all?
I recently saw a video on the making of Monty Python's The Holy Grail and the biggest reason the ending scene happened is that they just couldn't afford to have a large battle scene which the "cop out" was the solution to.
Was every element in that film design to be funny? The jokes wouldn't have landed without the context of the setting being accurate enough for the subversion to happen in the first place.
Nobody would probably see this comment, but I definitely agree with "limits breed creativity". I am a comic writer, and I used to have this issue where I would just fill the pages with useless fluff, and I would feel my stories walking in circles unable to fully progress, so I told myself "ok, you will now only have four pages per chapter. Do what you will, but you must have a complete story arc in just 4 pages". And I started to think about what I actually wrote, because I didn't have the luxury of space. Every panel, every line of dialogue needed to mean something, because if it didn't, it would just take space from something that did. That's why I say "Yes" one hundred times to limits. If you don't have natural limits to your work - create them. And that's what a lot of modern games don't understand, and why they start to wander. Why the feel watered down
I see it and I agree. I dabble in writing stories in my spare time, some fan fiction, some original stuff (mostly Fantasy) and some D&D campaigns. Core to all my work is always to look at creating the boundaries, the walls and ceilings for whatever powers will be involved in my story. In Fan Fiction I can look to whatever I am writing from. It may not always be stated explicitely, but looking at the application of powers or other parts of the world can often inform you. In paticular, the lack of something can often tell you just as much as its presence. Hypothetical example. You have a world populated by people who all have some kind of elemental power. Everyone has one power and only one power. Logically the fact that all people only have one power would probably mean that offspring of two people with different powers only inherit one of these powers and is not a hybrid of some kind as, assuming some amount of mingling between those with powers, if hybrids were a thing, one or more would have popped up by now. Now this is more a in story limit rather than a creation limit, but I feel they are just as important cause if you dont set those limits on the world then you might be tempted to just bend the rules if you write yourself into a corner...and suddenly the Fantastical but logical world you created starts to crumble.
Deep Rock Galactic is based completely on procedually generated Levels and its friggin amazing. Really feels like a whole planet full of caves, tunnels and glyphid hives.
That's because each room is actually designed mostly bespoke. They just string them together and procedurally generate some geometry and interactives inside those rooms. The actual gameplay loop is pretty tightly curated.
Deep Rock is almost entirely driven by gameplay and teamwork, means that the short falls of procedural generation with story/characters/plotlines doesn't apply at all thankfully
Sounds like I have to give this game another try after like 3 years or so, because gameplay was just as repetitive and boring as the generated worlds back then when it was rather new.
Yes, and is a fantastic example of the gameplay must also be good for procedural generated worlds to work. There are some absolute fantastic examples of it working really well, such as roguelikes etc.. but when someone uses it as their unique selling point, it's a massive red flag. Deep rock galactic has so much charm and other features that most players wouldn't even know it's procedurally generated because everything else is pretty good.
Deep rock galactic was one of the few games that procedural generation got right. Every cave felt and looked different, then you play for 50 hours and see that what you actually get are a random selection of repeating rooms, spaces, corridors and connectors. In a game about exploring and spelunking maps knowing your way around and where all the resources are removes half of the game
The thing with DRG though, is you'll never have the same experience twice. You might run into fundamentally the same rooms, but the moment you remember that you're in a cave system, you kinda have to remind yourself there's only so many ways for caves to form, and that helps cushion the idea of repetitive rooms, especially when your Engineer this round is trying to dig a hole straight to the bottom of the map and isn't helping with anything, but he digs a massive hole and you are compelled to go check it out, because fuck it, what's he actually doing? At that point the mission becomes secondary because players are using the sandbox that's available, and you're just curious as to why there's nothing quite like elsewhere, because in that game, the players you get along with you are what determine whether or not you're going to have a fun game, and you virtually never run into a hostile player who wants to ruin things because you're in a game that encourages positivity.
@@nuclearsimian3281those moments are the best thing about drg, the moments when the mission goes out the window in favor of yelling "we're rich!" Or pondering the orb. It's such a great design that it's addicting. Rock and Stone!
I think that games which take heavy advantage of three-dimensional content can benefit the most from procedural generation. Deep Rock Galactic does this very well, but Minecraft also does very well with its numerous biomes and things like cave systems to explore. With a third dimension added to procedural generation, it becomes more complex to render and piece together properly, but the result often can be spectacular. I would like to see more games taking advantage of that kind of thing. Starfield with some procedurally generated caves that are worthwhile to explore definitely would have been much better, at least.
I love DRG but the caves absolutely do not look or feel different. They are made different by the chaotic and dynamic gameplay usually by the end of a run, but when you first launch in, other than mission type, very little about the levels themselves is different
Most roguelikes in general are a good example of procedural generation...and that's why I love almost all of them even with the punishing difficulty that comes with some of them Starfield is a bad implementation of procedural generation...which reflects the title of this clip which the game is filled with (literally) a lot of "nothing"
I immediately thought of Subnautica when Josh was talking about this. Initially (in alpha/beta I think), it was going to be procedurally generated. However, when they added in the story, they chose to cut the procedural generation in favour of more curated, thought-out world design, something that worked out greatly in the game's favour (in my opinion, at least).
On one hand it gives an amazing first gameplay experience especially when you talk to other people and they have similar experiences as you such as the reaper leviathan by aurora, or the "multiple leviathan class creatures ahead, are you sure you want to continue". On the other hand, the game has 0 replayability and once you complete it, you can't help but feel like you want to play more but subnautica doesn't scratch that itch anymore. So you feel kinda empty for a time, try find similar games but nothing comes close to subnautica as it genuinely is the top of that game genre. So honestly it's just a difference of opinion and perspective. I don't think one type of game beats the other, but it depends on the individual and what they desire in the moment. I personally like a balance, I play a lot of grindy games (not always necessarily procedurally generated but not lore heavy either) and then some lore heavy games. So there is a need for both types of games to exist.
@MrTalithan strictly speaking, the fauna are procedurally generated by and large. There are rules, obviously, but it helps keep the world moving without blowing out your processor.
@@MyNameIsSalo "the game has 0 replayabilitiy" and yet most people I have talked with about the game have 2-4 playthroughs. To me it seems that on average it takes the player 2 playthroughs to really get familiar with the world, as you can easily miss a lot of stuff in the game world during your first playthrough and the second one becomes much more about just the exploration of the world since you already know what you have to do with the plot so you feel like you can take more time just looking around. And then there's the people who start new playthroughs just to try and build their first base on whatever biome so that they can look at leviathans through the windows.
One thing that shocks me in Starfield is if you're in a major town or city you can shoot a gun one inch from an NPC's head and get zero reaction, like GTA 3 had that figured out in 2001...
@@Pangora2They we’re too busy jacking off on their procedurally generated soup of a world and had no time to consider actual immersive role playing elements.
@@junojuno3745 Before Cyberpunk there was Anthem, Fallout 76, Watch Dogs, The Division.. etc… games that come out and are stable and polished at release are becoming the rare exception sadly
I was genuinely impressed by procedural generation back in Elder Scrolls Daggerfall. The idea that you were walking around on a landmass the size of a real country was so cool to teenage me back then. Of course, we always fast travelled and in the end it wasn't anything more than a gimmick.
It had potential to be so much more, but sadly they abandoned that and when they finally went back to heavy reliance on that technology, they do nothing to add substance with it. And personally I still prefer the way that world feels more real even while just using that fast travel system. It manages to facilitate immersion with it's systems rather than a dull feeling of emptiness
Being an IT guy for over a decade, and having been well into my career by the time "procedural generation" became a buzzword, has inured me to ever being impressed when a game uses that term in its marketing. However, to answer your question Josh, I will say that I have finished exceptionally few open-world map-marker driven RPGs, and the reason for that is because when I start a new one, I often feel that rush of excitement and anticipation: "Oh wow, this world is HUGE, there'll be so much to explore and discover and do, I can't wait!".....And then after scaling the fifth tower to uncover all the little markers on a new part of the map, and seeing that the number of "go here, collect/kill/mark X" icons dramatically overwhelms the number of actual bespoke side quest markers, I eventually come to the same realization every time--it's all padding to turn a 20 hour game into an 80 hour game, and not only is it boring and repetitive, but in most games all those little "to-do's" don't add anything to the gameplay, story, world, or characters I'm trying to enjoy in the game. If anything, they dilute it to a terrible degree. The only difference with procedural generation is that, as a developer, I know the nuts and bolts behind it, so I know to expect that feeling from the get-go, and consequently avoid any RPGs or adventure games with that as an advertised feature.
I would say that open world games really rely on gameplay and the world itself. Games like The Witcher 3 (I know that people complain about the combat, but that game has a beautiful world and the story is fantastic) and even Horizon Zero Dawn do this very well. They are not procedurally generated (as far as I know at least) but they are great open world rpg games. No Man's Sky has improved a lot but the procedural generation of it is hardly even the part of the game that people really care about. Yes you do get a bunch of interesting planets out of it, but for the most part the only difference is some color filter at some point
I think procedural generation has it's place, and can be impressive in right spots. But can't work everywhere. I think for an RPG, it's generally a bad idea. Unless in specific circumstances or sections. Like with ARPG's. But I would add that procedural generation can work wonders in certain strategy games. Civilization, Age of Empires, Heroes of Might and Magic (random maps), Stellaris etc. work very well with "endless gameplay" and procedural generation. And they can easily be 1000 hour games, where all gameplay feels tight. But ofc - those are all games where - a single "session" is not 1000 hours. It's only 20 min to couple of hours long. Some, like Stellaris can have longer sessions, but it's still not more than a dozen hours per. Similar with Roguelites. Procedural generation makes sure every try or every session feels a bit different. But a classic RPG, like Skyrim or Baldur's Gate - you aren't resetting your game every hour. You play a single session that is 20, 50 or 100 hours long. And if you stretch this too much, by adding procedural generation, so that experience lasts 1000+ hours, it will feel "empty and pointless". I would add that I think procedural generation could work in those games, by only applying it for a few sections - like dungeons in Skyrim for example. If they were procedurally generated, it wouldn't increase the time it takes you to finish the game. But it would make every time you restart the game, feel a bit different. Thus giving a bit more re-playability.
@Voidstroyer Exactly. And to be clear, I have no problem with devs using procedural generation for stuff like, say, building the base geography for the world map, or even populating that map with certain features like forests, etc. But the actual locations and wildlife and characters and quests, etc, down to the prop models, should be placed in a bespoke fashion by a designer with intentionality and purpose, otherwise players will feel the meaninglessness of those elements when they play the game, even if they don't know enough to identify why those things are clashing with their suspension of disbelief.
@Member_zero Yup! That's why I specified adventure games and RPGs. Obviously procedural generation is a fantastic tool for stuff like roguelikes/lites, as that's a genre that would probably not exist as it does without it.
His question about not wanting to play a game if its 500 hours because you don't have the time. It doesn't apply to myself. I do have the time. My problem is that I can never stick with a singleplayer game for longer than maybe 20h. Unless its a multiplayer game where I keep getting engaged by other players, I just lose interest.
I get what Josh means about how Nothing can be impressive in certain circumstances. The Glowing Sea in Fallout 4 was like.... 90% just featureless wasteland, but man did it nail the atmosphere of an ungodly irradiated hellhole so hostile to human life that it may as well have been an alien planet. Even if it wasn't actually dangerous to an even moderately prepared player - and let's face it, it wasn't - it was just so damn cool.
@@steaketc2476the only place Bethesda isnt doing just the minimum is the map design (pre starfield, even FO76s map is fun to explore most of the time)
Honestly it fucking did, I specifically remember the weekend I first played it, I woke up around like 6pm (nightshift worker) and played through an entire questline, made some money and eventually ended up with a ship at the stats I wanted and finally took a break, only to notice the sun was still fucking up outside. It had been like three fucking hours and I felt like I had been playing for like half the night already. If fun and enjoyable games suck away my free time, then that should have been the biggest red flag that I was playing a very boring, checklist of a game that isn't engaging me at all. I was just engrossed in the novelty and hype of a new game, a space game at that, which have even more sway over me.
I just walking around in one small place for Baldur's Gate 3, barely progressing to a quest, and it takes me somehow 6 hours, and it's still puzzles me why. Going on to one Skyrim dungeon for an worthless RNG loots, and the dungeon takes 30 minute to slog through. Doesn't want to bother buying or playing Starfield. Because afraid that the crappy old Bethesda formula is still there.
It's a funny point you make about procedural generation. I once worked on an RPG project where, as part of the setting's lore, there was a massive forbidden zone to the north--a blasted hellscape chock full of dangerous creatures, but nothing else. There was no xp to be had. No quests. No loot. Just an endless expanse of procedurally generated wastes and deadly monsters. In our case, we actually used procedural generation to discourage exploration. The entire zone was completely impossible to map out and completely pointless to explore.
Yet, there would be people who explore this area nonetheless. I recall a buddy of mine playing Sacred 3, a subpar game by all accounts, especially compared to Sacred 2. During one mission, in which you’re expected to charge a rune while enemies attack you, he spent hours on it just because the enemies were endless. You were supposed to kill enough to charge the rune, use it, and move on. He sat there fighting these enemies for hours, gaining nothing but small amounts of gold, just to see if there was an actual end to them. Apparently not, but he had to go to work so he had to stop.
@@user-os7ec4dm8x it's like you're completely missing the point of procedural generation. they got to put more time into handcrafting the parts of the world they wanted the player to explore, without using an invisible wall.
I absolutely agree. I honestly get nervous when a game I am excited for has a main selling point of "this is BIG game." Right when they start doing this im like "O God here we go again." I think that the problem with procedural generation is that the tech is simply not there yet. People are not dumb (for the most part), we can pick up on patterns pretty quickly. The problem is that the minute you do pick up on it is the minute boredom starts to creep in. This is exacerbated 10x when the game relies on this technology way too much.
It isn't that the tech is there yet, because it's there - it's been there since at least Daggerfall, most of its landscape is procedurally generated. The problem is that it isn't actually very good, and the few people who actually sell it are more interested in either the math behind the coding or the illusion of endless possibility. By contrast, devs and publishers like it because it means less time spent on coding or asset creation, which to them means more money saved that they can budget elsewhere (or magically put back in their own pockets). It really is a lose/lose situation for most gamers.
The core issue here is that while it's easy enough to fill a game with tons of stuff, it requires actual design to make it good. It's possible to procedural generation right, provided you actively involve competent designers in the process. The simple solution that a lot of roguelikes have gone for is to build maps out of preset tiles, with the tiles being designed elements. The obvious issue with that solution is that players will very easily spot repeating tiles, so you still aren't getting a lot of mileage out of the elements that were made. Going a bit further than this means building design logic into the algorithm which can result in a less repetitive environment, but the player will still likely never see anything that wasn't specifically accounted for in the design in any case. No matter what way you go about it, you will not wow players with hundreds of hours worth of remixed elements. The only way to keep players entertained for that long is for the gameplay to actually be good. More often than not the only real value the player gets out of procedural generation is not knowing what's ahead. It adds replay value if the game was worth playing the first time but even then, once the player learns the patterns it's no different than repeating any other game.
They're just not using it right. Hand crafted continents and landscapes followed by a few procedural generated dungeons for when the hand crafted dungeons have been plunged. I mean, it works for Diablo 2.
The simple thing is, Bethesda does not know how to make the best use of Proceedual Generation tech and Starfield shows that very nicely. Another way of using Proceedual Generation Tech and in a more correct manner, is the way that CIG is using it for Star Citizen, they create a planet with proceedual gen tech within certain perameters and after it has been created, they let their artists work on said planet to bring life to the planet, think of it as, painting on a newly created Canvas. Starfields is.. Copy-pasta proceedual generated "stuff".. The way the two companies are using the tech, is miiiiles apart and Bethesda has no idea to to properly utilize it!
@@kerianhalcyon2769 it definitely hasn't been there since daggerfall lmao. that game was utter garbage, and its procgen was some of the worst that's ever existed.
Yakuza impressed me with its map. It's absolutely tiny but they make it feel different in each game. Seeing the 80's version of it, then early 2000's and then evolving throughout 2010's. Seeing the same noodle shop stay in place and not changing, you feel nostalgia for a place that never existed.
The most amazing example of procedural generation is Dwarf Fortress. The stories that its capable of generating is absolutely amazing. Humans are really good at pattern recognition, so people start really noticing the repetition that comes out of procedural generation and everything starts feeling the same. As a developer you need to put in an exceedingly large amount of work into making it so that this doesn't happen in a players' playthough, but honestly that's probably more work than actually curating the entire experience by hand.
The thing with Dwarf Fortress and Rimworld is that the procedural generation in a whole different scope from say... No Mans Sky or Starfields procedural generation. The colony sim style games work their magic by using procedural generation not just on maps but 'events'. Starfield's procedural generation is 'here's a randomized map with randomized stuff on the map.' Rimworld's procedural generation is 'here's a randomized map with randomized stuff on it, and here's a randomized weather event, and heres a randomized raid of randomized attackers, and here's a randomized illness on one of your pawns. Etc etc etc.' And on top of all that, the 'randomization' is actually curated by a storyteller in Rimworlds case. That cascading effect of overlapping all these differing procedural generated things is what earns these games the name 'story generators.' And its how they avoid the 'mile wide, inch deep' effect we see in games like Starfield.
@@FTORanger Bro, you're in for a trip with Dwarf Fortress. It's pretty overwhelming for a lot of people at first, and is one of those games you rely on a wiki for. If you like the concept of Dwarf and can't click with it fully after playing some, I'd 100% recommend trying Rimworld out after that if you're in love with colony sim games. Rimworld is a great entry to the beast that is Dwarf. Same gameplay concept, different themes, narrowed down a bit, and has in game tutorials and tips lol. Just, if you don't know how to play these types of games, losing is part of it. You WILL lose, so don't feel bad if you do! It's all part of learning how to play lol.
Also a very notable thing about roguelike procederal generation: very often they have set building blocks which are remixed and connected in different ways, but do have effort put into their design.
@@SawGudman I wouldn't even really call Path of Exile PG, there are MS paints-like maps and it always picks one of those and just makes the terrain on them slightly random. It is PG, but it's the smallest amount of PG you can use before you're just hand crafting maps.
procedual generation is incredible, when it has restrains, guides and coherence. players dont need infinite content, players need the surprise and a meaning on these suprises.
Honestly the best take in this comment section. There's a time and a place for it, and when its done right, it can be a very useful way to enrich a game.
Fans just expected Bethesda to EXPAND on things they used to do, not dumb them down even further. People wanted procedurally generated planets and dungeons akin to Daggerfall. Yet Starfield just plops the exact same prefabs on barren landscapes. People wanted NPCs to have daily schedules like in Oblivion. What does Bethesda do? Removes daily schedules from NPCs entirely. People wanted faction choices to matter like in FO4/Morrowind, but yet again we get Skyrim's "you can do everything in one go". It's mind-blowing.
I loved the concept procedurally generated Levels but after seeing them for years only a few of them stand out and work. Mainly in the rouge lite genre.
The only reason this will work is when Creation Engine comes out and modders spend two years filling the content out so you don't encounter the same 'secret' mining facility or lab for the 15th time.
Yeah for roguelikes it works because repetition is the intended point. In games like starfield, there's no reason why repetition and repetitive environments should ever be viewed as a good thing.
That is because procgen cannot stand out by definition. It can only remix assets that only exist, and in most games that means you don't really register the actual layouts of said assets, you notice the assets themselves.
Procedural generation isn't a bad idea, even in the "bad games" Josh mentioned. It's just a bad execution. The whole concept is basically "done dirty" by how it is used: a simplified randomization of pre-created assets. A few different tables at best like "biome", "in-door assets", "out-door assets", "loot table", "enemy table"... mb something like "miscellaneous". So it obviously generates dead-on-arrival, sterile environments. And it isn't that hard to fix. Add more conditions, add more criteria for picking assets, add memorization so that areas won't be a copy of each other with 1 element changed, add statistics analysis so that your auto-generated level would be on par with manually crafted ones. I truly believe that it isn't that hard, really. It would just take time. More time. Probably more people involved. And more testing to polish it. And all of those things mean "MONEY": time is money, adding more people to a project means money, testing/polishing something requires money... And a company as a whole doesn't really need a "good product", they need a product they can "sell". So instead of a tool (talking about procedural generation as a tool here) that could makes game GOOD, they just use it as a tool that allows them to SAVE TIME (=SAVE MONEY). That's really all there is to it.
Dwarf Fortress is the worst possible example if you're trying to argue that it's "not that hard". That game has been in development for almost 20 years, and it is massively complex, with dozens of incredibly intricate systems interacting in ways that the devs have to make sure doesn't break the game. That is a very hard game to make, and would only get harder with a large team because of the coordination required between disparate systems. In general, good proc gen is so much harder than "just add more conditions lul".
In terms of procedural generation approaching some kind of ideal that AAAs could never touch, I always think about Dwarf Fortress. Its whole world and all the characters and the entire written and unwritten history of everything and everyone are generated. In the game's fortress mode, this mostly affects you with the megabeasts that occasionally invade your fortress and the visitors that loiter in your taverns, the personal histories and skills/knowledge of the migrants who move in and become your citizens. But also with things like your civilization's technical knowledge (how much of the invisible "tech tree" has been opened up already, with scholars and a library needed to get more), news from trade caravans about recent events in the world, dances and songs and the instruments needed for them created over the years by bards and musical troupes, engravings on walls pulling directly from historical events. Then in the adventure mode (which was temporarily cut from the game's steam release but will come back some day, maybe next year), it was a lot more jank, but as your adventurer party roams the world, you could ask any random character out in the world about any character, location or rare artifact your party has ever heard of, and usually they probably won't know what you're talking about, but other times they could point you in the right direction and help you find your way with the (again, janky) quests. It always fell short with how limited the dialogue from the npcs was and is, though, unfortunately, as that's been one of the game's weaker spots for a while. You can ask a dizzying number of pre-formatted questions with the subject swapped out, but what you get in return often just ended up being "it is inevitable" or one of the other go-to phrases. There's definitely so many more years and updates needed before most of DF's features feel better, more cohesive, more playable, since their ultimate design goals are ambition incarnate, but the devs still put a hell of a lot of work into the game's systems as it is, and plan to put in a hell of a lot more, and it's magnitudes more time and effort (and money) than what any AAA devs or their shareholders would ever be willing to invest in one title. But it gives me hope for better and more nuanced implementations of procedural generation in games, just by the merit of it existing, even if it is a one of a kind game.
When you think about it, the entire real universe is procedurally generated - just with incredibly complex conditions and rulesets: given identical conditions, thr ro k of a cave system will erode the same way every time. But of course they are not identical conditions: the rock has different makeup; the water varies in quantity, and ph; the caves all started at differing times... Planets and suns form by the same rules - they couldn't be more different.
@kekker_3362 but: A) it was still made; B) people loved it. Of course making such a system isn't as easy as to describe it in a few sentences in UA-cam comment section. I was not implying that)) But it's also not a ground-breaking invention, that requires dozens of geniuses to achieve. And also saying that it's harder because of number of people involved is really funny, considering how big game-developing teams are nowadays. Have you ever watched game's end-screens? Hundreds of names! Adding a few more - and in most cases you don't even need to add, because they would probably be the same people already working on the game! - won't be anything new. Coordination is a task - I can agree on that, - but it is already tackled. You can't use it as an excuse for not doing something. It's kind of more ridiculous than spending more money on a project.
I like how Subnautica mixes procedural generation and design. The map and most landmarks are fixed points, and some creatures are placed with pathing and AI. However, the resources and most of the creatures are procedurally generated based on the biome. It makes each playthrough different and keeps the world feeling alive, and it doesn't mess with the progression or shatter immersion.
The one thing that really throws me for a loop in Subnautica is how the life pod spawns at different points in the safe shallows on each playthrough. I'm always lost at the very start
Oh, I didn't even know that, and I love that game! The thing is, the game is so huge, yet its all made by actual people. So I didn't even notice there was a any procedural generation going on at all.
@@tenyokensekia8088 Because Procedural generation isn't inherently bad. Some of the best Dungeon crawlers are procedurally generated. You can make it work, it's just that Bethesda decided to use procedurally generation to replace everything. I'd love a modern RPG that did what Diablo 2 did. The procedural generation is not the problem, it's just the people, and it always has been just the people.
also the beauty of nothing comes when you don’t really expect it or you know how you got there. my problem with starfield is it’s like knowingly taking a loading screen to just sit in the middle of nowhere. it takes away the magic and doesn’t feel like you stumbled across this vast landscape you just loaded into one
Since Starfield I decided I'd skip it and just play Morrowind. Vastly smaller, but my previous playthroughs I avoided whole areas of the map and whole quest lines. Its interesting and I have saved a game out in the wilderness and lacked tools to get an NPC escort home. Also Kerbal Space Program is a game about dead empty planets with nothing there. Its amazing once you get into it and start building ships that can reach new worlds. They made the mechanics of travel the key, and its an exploration game. Finding anything weird at all is something you run to take screenshots of and show a friend.
I think part of the reason studios go for bigger quantity goals is to have “objective” metrics to brag about. They love to say the numbers of planets, the area of the map, the line count of the dialogue (dialogue is something starfield specifically bragged about in ads)
This would've been VERY impressive in 2010, but in 2023 we've already been fooled enough times by massive worlds with nothing to do in them. I honestly can't believe there were people excited to hear about the 1000 planets.
It's also a fantastic fit if you wanna sell a subscription service like Gamepass and cannot rely on the good old 70 bucks per game sold on release kind of business model anymore. Probably why the new Forza game has you level up each of the 500+ cars individually. I hope the whole subscription crap isn't gonna result in the next race to the bottom akin to what happened with the mobile games market.
I think they're just behind the times a bit. There was a time when open-world + sandbox + big was a massive novelty, a technical achievement if you could pull it off, and something players thought they wanted (as opposed to being 'on-rails', rooms+corridors etc). And people did. Then they saw that maximalist interpretations of that aren't fun and aren't as good as they thought they would be, so they've settled on wanting something smaller, with a good, hand-made, tight central plot, some curated side quests and plot lines and a sandbox for those who love the world and want to go off and live in a tree surrounded by modded-in s-xbots. The execs just haven't caught up. Atomic Heart is a good case study. It was going to be a big open world sandbox originally (with a plot), but pivoted to a much tighter scope and benefited from it. There were a few grumbles in reviews about the wider world not being as big or populated as they expected based on some early discussions, but no one really cared enough about that to take more than a point off the rating. If you can't make an engaging big world, don't waste the money and time because players will not appreciate big for the sake of big in 2023.
@@CidGuerreiro1234 Exactly, the game is outdated, despite coming out this year. It makes sense when you remember that Starfield is built in the same engine Morrowind was built in.
Which of these two games has more quests? Game A: 1) Speak to the townspeople to gather information and make a plan. 2) Gather items you need for your fancy plan from three different enemy types. 3) Carry out your plan to raid a big enemy base and defeat the big bad within. Game B: 1) Deliver 6 wolf pelts to a leatherworker. 2) Deliver 8 fire shards to a blacksmith. 3) Deliver 12 spider threads to a weaver. 4) Deliver 5 spearheads to the general. 5) Deliver a baker's dozen eggs to a chef. The kneejerk answer is that they go 3v5, but a careful inspection shows that Game A has three quests, and Game B only has one. Even if it had a thousand such quests, it would STILL only have one quest. That's what these game devs aren't seeing. And all that's not even considering quality, just quantity. A game with a thousand quests loses to a game with 3, on QUANTITY of quests. Quality is judged separately, but history has shown that the 3-quest game usually wins on that front too. Josh said he'd prefer a well-made 10-hour game over the 1000-hour procedurally generated one. I'd go even farther and say I'd prefer a well-made 10 MINUTE game. Which I can back up with a history of playing short little flash games on old sites like Armor Games, Kongregate, and the like.
Nah mate, my knee jerk is that it's 3v0. Deliveries aren't quests, they're side jobs. Ways to grind up a little extra cash, or your first minor task before your ready for the big leagues. If fetching some materials is a quest, then the word has lost all meaning.
@@lawaern3474 That's another way to look at it. Personally I don't mind being generous and calling it 1, especially since that leads to "at my most generous that's still atrocious". In moderation and spread out between more awesome moments, the rare kill-x-things quests don't bother me at all. But I also completely get it if even that much bothers other people.
@@riluna3695 They don't bother me too much either, heck one my happiest gaming experiences (gothic ii) had a kill-x quest at the heart of it. I just dislike the gaming term "quest" being used for anything an npc asks you to do. When I went out to get some bread from woollies this morning, it wasn't a quest. It was an errand. But when I crossed Gielnor to gather rare and absurd ingredients for One Small Favor, it was a QUEST. But, I firmly believe errands and other minor tasks have a place in adventure rpgs. But we play these games for the quests, not the errands. Calling them quests just gives game devs excuses to pad their advertised numbers.
@@lawaern3474 Ah, I see what you mean. Yeah, I agree completely. Best to call it what it really is. That also reminds me of an old favorite cartoon :D "So it's not really a mission, it's more like an errand." "Well missions are just errands that....involve death." "So like...a funeral...would be a mission." "...You think funerals are an errand?"
I'm a senior game designer who's been working in the industry for over 8 years, and I can confirm your suspicion that those who advocate for procedural generation are more often than not directors or executives with little understanding of what provides genuine value in a video game as a product.
This thing about limitations is something i always talked about. So many awesome things in old games were the result of technological limitations of the time and the need to find a creative way around them. Today they have the technology to run wild with any crazy idea but they can't come up with ideas because there is nothing forcing them to think outside the box. The easier the work is, the more formulaic the product
Nah, still happens in indie games. Public traded companies just have absolute inhumane monsters as execs and suck all fun and innovation out of video games. Just imagine what a studio's burn rate is with 400 plus employees. They have a hard time justifying taking risks. So they just copy what worked in the past.
@@1IGG And the other part that a lot of folks don't like talking about is that consumers don't like risk either. Even if a change to a franchise is good for it's health in the long run.
@@1IGG But indie games do have limits. Limits in their budget go super far. They often can't have a huge team of artists to make something on par with AAA so they are limited to something more stylized. The budget for sound effects will limit their ability to compete with Call of Duty and Doom 2016 for satisfying sound effects and gun feedback. And engine + creative limitations are still a thing, you don't have teams of hundreds of people to figure out *how* to do something you may want to, so you gotta figure out a workaround. The only limitations it feels like retro devs had that moder. indies don't is hardware limitations. Otherwise, Super Metroid and Castlevania SotN weren't made by any more than 12-30 people depending on how you count musicians.
I recently left university with my masters and started at an executive level in a company (not gaming) and I see what you mentioned in both my colleagues and higher ups as well as finding it deeply frustrating It’s all about making numbers go up at the expense of a good experience that actually makes sense and engages the customer
@@yonosepe7491 This is a meaningless comment. Lots of companies make *good* products while covering costs. It's greed pushing these companies to forgo average customer experience to chase ever-increasing profits. How are you having trouble "covering costs" when you report record profit to your shareholders? Those are diametrically opposed concepts. Bethesda in this case dumped hundreds of millions of dollars into Starfield. It wasn't a lack of funding or cost covering issues. It was poorly using hundreds of millions of dollar and *years* of workhours mismanaged because they bought their own hype and several senior members working on these games have a mentality that because Skyrim and its 12 versions or whatever sold amazingly that their ideas are above reproach and criticism. They were so wrapped up in wanting it to be the absolute smash hit like Skyrim that they just made Skyrim again but worse, and now Starfield's cultural legacy will be highlighting how truly out of touch the leadership at Bethesda Game Studios has become. They don't want to *work* for the accolades and sales figures, partially because of Todd's trauma with the Morrowind crunch as they teetered on bankruptcy, and partially because many of the senior members think they deserve it, despite their formula being *decades* old at this point and pretty tired on its own. There's a reason more people have played modded Skyrim than people who have finished the main quest lol. At this point they're better at just making models and worlds that feel good to wander around in. Nobody is talking about an inability to afford the work needed to produce something good. That's not even a factor in this equation. The number couldn't be going up if the costs were outweighing the profits.
Starfield somehow manages to be a game about exploring space that somehow manages to lose the wanderlust. Once you have played 3 hours you’ve played it all.
Agreed. No man's sky is the only good procedurally generated space exploration game. Starfield was outdated by a decade on release. But people will still meat-ride it due to sunk-cost fallacy and wool-covered eyes.
I put 100 hours into Starfield purely due to the sunk cost fallacy. I tried to convince myself that the bits of fun i was having were enough when they clearly weren't. Meanwhile, i beat all three endings of Armored Core 6 in the same time and still think about how much fun i had playing it
446h on bg3 and currently i'm still on 2nd playthrough. this is truly mindblowing. there are so many to discover even on my second playthrough, and I'm on the "good" side on both. the game is quite buggy in Act 3 and Gale is accusing me of making an agreement with a person i had yet to meet, but the game as a whole is still incredibly well made. it is a gigantic flawless gem with a few scratches that they are still polishing.
I had the same issue in BG3. Also for some reason I have an option to ask a companion what happened to our romance even though the romance is full on ongoing. Minor bugs and AI movement aside, the game is really deep and i love it. Just like a great RPG is meant to be.
Yea a couple of times my companions seem to know more than I do but that happens in real life too so I let that bug pass. Awesome game. It’s turning into my fav western turn based game of all time
The GDC presentation for Into the Breach was eye-opening to me: that game has no procedural generated levels, despite being essentially a rogue-like puzzle game which, on the surface, would be perfect. Instead, they realized it would be faster to make hundreds of challenging and fair levels themselves, rather than perfect an algorithm that might not work (or probably not be fun). That same reasoning could be beneficial to a lot of games.
@@santiagorios2083 I didn't realize that for Hades, but now that you mention it: the Binding of Isaac is set too. It proc gens the map layout based on simple rules, but the rooms themselves are hand-built (with rewards scrambled by seed). FTL and Slay the Spire also only use limited proc gen.
@@santiagorios2083Hades is a game that feels like devs play games not as work but as passion and hobby. All rooms, despite being able to be crazy hard, are fair. Even the shittier ones.
The problem with how a lot of AAA devs use Procedural Generation, and Bethesda are kings of this, is that they use it for the sake of it. It’s a marketing bullet point, rather than part of the core game design equation. You look at something like Dead Cells where part of the challenge is not being able to memorise the level layouts because it changes every time you die, and you get a design that was built around Procedural Generation rather than one that has it bolted on to the side. To make a comparison to a more similar game: No Man’s Sky. It’s a game where every environment is made with PG, but the entire point of the game is exploring these endless new planets, and each one is a complete planet with its own biome and flora and fauna. It was a core part of the design, and the game is better for it because practically nobody has seen the same planets as you. NMS’ PG is used to give you an endless amount of exploration and discovery. Then you have Starfield where it’s basically used as a clunkier and less interesting version of Skyrim’s caves because it was just an easier way to add a large amount of content to the game rather than making content that’s worth experiencing for the experience itself. If Procedural Generation isn’t a core part of the game design, then what you generate with it has to be something incredibly entertaining or it’s going to feel like an all-you-can-eat buffet of flavourless slop.
I don’t think any other game filled me with as much awe and wonder as consistently as Elden Ring did except for maybe Outer Wilds (but that is a much shorter game).
@@mechanomics2649 Is that why it made Fromsoft much more popular than they were ? Elden Ring got beyond mainstream and let me tell you the amount of people who played it who hadn't heard of fromsoft is absolutely enormous. It is an insanely high-quality game and them re-using a boss 3 times won't change that.
I think procedural generation can work occasionally. Hades and Remnant: From the Ashes both had a bunch of curated rooms that would get laid out for you. In Hades, it’s before each run. In Remnant, when you started the game it was procedurally generated so you’d have a unique map to someone else, and that was a cool idea, but every area was made and then connected to each other. You’d have a boss fight in one play through that wouldn’t show up in another and that was cool. But in Starfield it’s just a bunch of empty planets for what?
Looks like someone hasn't played Starfield. Look tbf its a bethsoft game, its fallout and skyrim elements put into a space genre game. But still your comment proves you haven't played the game. So cope.
I dunno about Remnant, but Hades is not procedural generation. There is a preset selection of assets for each level, a preset layout and preset potential enemy spawns. It randomly selects from the assets available, it doesnt just CREATE new layouts from the assets.
Sometimes the 'nothing' can be used for great effect, it can be Poignant and a breather for the player. The long ladder from snake eater, the quiet paths in dark souks that connect one area to another, the open areas in shadow of the colossus to emphasise how big the boss's are. But big open nothing for the sake of nothing is just pointless.
I think it would be beneficial to distinguish between a randomizer (Hades) and procedural generation (Starfield landscapes). When you take a big pile of handcrafted elements and draw a handful out of the bag and let the computer string them together randomly thats for of a randomizer than procedural generation... procedural generation creates topography, forests, puts in animals and sometimes enemies.
Still, making the distinction does not tell whether the game is good or not. Minecraft and Terraria's worlds are fully procedurally generated but they are great games and they wouldn't even really work without the procedural generation. But that does not mean it works in games like Starfield.
Yes, and if it was implemented on Skyrim it was definitely checked and edited by hand, which can take longer, so possibly not even used. The Starfield moonscapes are obviously generated - planets and moons obviously can't be made by hand. Starfield has large hand-crafted bases, but the 'open world' beyond is really just tacked on, as it's expected, but it's almost irrelevant, hence making moon buggies was not worth it for them, or too difficult. They couldn't have Skyrim's great explorable open world and space travel to hundreds of planets. Considering how dull the space travel is, they really should've had one huge planet, with maybe a moon, and an agile spaceship to fly and land to explore (a bit like #HaloInfinite, but far bigger, and better spacecraft than the Banshee) - but that's all technically beyond Bethesda, who'd rather use the processor power on the 1000 looted items you have in your pockets or have scattered everywhere.
@@goatlps Having only one planet would have drastically limited the location diversity, biome diversity and the worldbuilding. Also it doesn't have the space exploration vibe, you won't be discovering new systems, you won't be finding beautiful planets to build outposts... The space travel is dull ? Even in No Man's Sky you need to "fast travel" between solar system.
I was so impressed with all the marketing around Skyrim's "Radiant Quests". I can remember being in Skype and Teamspeak with all the mates I played with online, we were all so hyped about it. Infinite quests, we could just keep playing FOREVER!!!, no need to worry about running out of things to do. Then we actually got our hands on the game, and while still a revolutionary game for it's time, all talk of "Radiant Quests" being the next big thing died immediately. It almost seems that every couple of years it gets brought back up again, that technology has progressed and "this time it will be different". But nothing I have seen has changed my opinion on em.
It's just an another ridiculous example of the way Bethesda are stuck in the past and coasting off past success to the absolute extreme. They're repeating the same misguided adoration for procedural generation long after everybody has seen it for what it is.
thing is Radiant Quests arent "new" in Elder Scrolls actual Radiant Quests started in Daggerfall.... and... Daggerfall being Daggerfall.... YEAH... some people have even gone about speaking that Starfield uses Daggerfall as its precedural generation process which may or may not be far from the truth as there are only 10-13 dungeons in that game that are hand crafted the rest are all procedural generation with expected jank
There's a few marketing buzz words that just seem to enthral gamers for no logical reason. A big one I can also think of is a game having a "10-year plan" All you have to do is look at GTA v to realize that maybe a 10-year plan isn't actually the best thing
Yeah, it's bizarre, because Morrowind was such a hit precisely because it got rid of Daggerfall's procedural generation, and everything was done by hand!
I told a buddy I did everything in skyrim and he goes "there's literally infinite quests" and like, technically yes, but it's really just the same thing over and over again.
5:50 YES, absolutely! I used to become incredibly hyped whenever a new "massive sprawling open-world RPG" was introduced but I learned my lesson. Nowadays, I try to temper my expectations and adapt the "wait and see" approach. I'm still interested if I see these types of games being announced, and I'll follow up on first news and reviews, but I wait until the game is released and then some more to see the first user reviews before I make up my mind whether I'll buy/play it or not.
Daggerfall is a really interesting game in procedural generation. The game does repeat random quests insanely and the dungeons were randomly seeded when the game was made, minus main quest dungeons, but still the game has this really addicting gameplay loop that makes it satisfying to play.
I tried playing the Unreal version of Daggerfall a few times. I really would love if they could make a game with modern graphics and controls but keep everything gameplay-wise that made Daggerfall and Morrowind great. I wish Bungie had done that with Marathon as well. The Marathon universe with Halo Infinite graphics and gameplay would be a riot.
Indeed, the gameplay makes Daggerfall work. Imagine this: port Skyrim or Starfield BACK to Daggerfall graphics. It will be clear how much gameplay was lost.
The point of procedural generation imo is to give a sense of scale quite easily without needing to produce alot. Daggerfall does this pretty good because the worlds as a whole are massive but the important elements are all hand crafted. But then again small worlds but with more detail is better than bigger worlds. For example the yakuza games have small open world but very detailed and each street corner is filled with life (kamurocho especially so is tiny but high quality).
@@OnlyGettyYou can say that of many many games. I don't get why people are so mad at Starfield. I mean, I understand the disappointment, I am disappointed myself, but I'm just trying to enjoy the good bits instead of shitting on the game at every chance I get. And there's A LOT of good stuff in Starfield. I am 50 hours in and I landed on maybe two planets that were not related to a quest. I barely scratched the main campaign because of the amount faction and side quests I got involved in. I just avoid the boring stuff. Nobody is forcing you to explore every single cave and talk to every single NPC that says nothing but useless boring dialogues.
@@thebaffman4898 Let me help you with that. Compare the game to Skyrim. How much has changed between the two games? The answer is: Surprisingly little. Starfield isn't a space game. It's a Bethesda game that takes place in space. Or maybe it would be more appropriate to say, Starfield isn't a game. It's a game engine with a story attached to it. Like Skyrim, life will be brought to the game through mods.
I said in our whatsapp in June that starfield would be a mile wide, but an inch deep. A couple of weeks after release, we had a news article with that exact headline. Every time a dev keeps pushing the massive scope of a game world it ends up the same way
1:22 I think Josh is wrong here. By now gamers might have somewhat caught on, but I still remember how massive the hype for No Mans Sky was, how often I was told that it’s basically infinite and thus going to be the best game ever made. Or even the “hour per dollar” calculation, as in how many hours you can play a game and comparing that 1:1 to money spent. A massive amount of people believed this to be a good way to judge if a game purchase is worth it, many still do. So it’s not just executives who want these massive games with slop content, a lot of gamers do too.
A great use of procedural generation is in Dwarf Fortress. The reason it works is how deep it ll goes, as the end goal of the devs is to simulate all of reality. For example, when a necromancer attacks you, it will not be some random enemy, it will be a person with a history, what they life was like and what lead them to learning the secrets of life and death, their rise to power and eventually the moment they attacked your fortress. If accompanied by, for example, and undead dragon, the beast will also have history of what they did before being struck down and eventually being revived by the necromancer years later. This depth is also applied to civilizations as a whole, scholars, heroes, artifacts and many other things. In a different game it would likely be a randomly generated necromancer, with a randomly generated dragon by their side, with no further depth.
Procedural generation works best in games that are designed around it, like Deep Rock Galactic and Dwarf Fortress (funny that both games are about dwarves). Bethesda, on the other hand, used procedural generation as a tool to additionally inflate their world into a bloat of blandness.
@@sujimayne Yep, procedural generation has it's uses and now that executives in these big companies finally found out it exists, they use it for things it is not designed for in order to inflate their games and save on time and work it would normally take. It's used to generate worlds/levels in certain games for a reason. Those games use it more as a backdrop rather than a deep world to interact with, making it a perfect solution. If it's used for characters, lore, etc, it needs to have a lot of depth, which requires much more effort to accomplish, than doing it the regular way.
The moment I realized procedural generated quests where a handful of nothing was in Oblivion after I became the Listener for the Dark Brotherhood and started getting those weekly meaningless contracts.
Speaking of limitations. I have two examples. One recent and one a bit older. Dead Island 2 does not use traditional shadows. It is a trick. It saves a ton of resources by making the shadows feel and look like they are true shadows casted by objects but they are not, they are literally drawn. It's so good that only in specific places and very specific camera angles can you actually tell. The other one is Xenoblade Chronicles 3. Arguably the most beautiful game on switch. And there it is. Switch. It manages to look ridiculously good while being limited by hardware. It doesn't render all the flora around you at all times and saves resources that way but still gives you the impression that everything is being animated all the time. It holds stable 30 fps while 7 of your characters are doing ridiculous over the top stunts fighting 5+ enemies doing equally ridiculous stunts full of particle effects. We need limits to find our true abilities.
I completely agree here. One of my guilty pleasures, I suppose, are the Assassin's Creed games, but the plot is just a small percentage of the entire game. I played Odyssey for like 150 hours and spent like three quarters of the total time clearing out the map that is littered with the same two or three basic activities. I'm so done with games claiming to be hundreds upon hundreds of hours in playtime, but once you play them you quickly realise they only achieve that time by having a massive, mostly empty world with about three or four activities, which are being repeated time and time again.
Deep rock galactic does procedural generation really well. They put the generation on a track with some set rooms with random variants in actual geography and random gameplay elements. You eventually notice repeating room designs but it takes a WHILE. Like easily 50-100 hours before you notice constantly if you've been to this room before. Rain world also uses this tool well. They randomly generate some creature animations with some animations being set. It works well because the creatures don't resemble real animals so our frame of reference for how they should move is limited. It nails that feeling of a living world by far the best.
Yeah, procedural generation is just a technique, it's not good or bad, it depends on how it's used. There are a lot of games that uses procedural generation very well, you can't just say "procedural generation sucks because in Starfield in generates a lot of nothing".
Yeah, the real question is what the procedural generation is used for, and have good is the algorithm. Minecraft is a good example of a great algorithm, to this day I enjoy just strolling through and seeing what landscapes generate.
@@mrosskne Dwarf Fortress, Minecraft, Terraria, The Binding of Isaac, XCOM 2, Dead Cells, Hades, Rimworld, Deep Rock Galactic, these are all great games that use procedural generation in some way.
I remember josh going over this exact thing in his Castlevania review, where the limitations of the engine are what pushed them to get the most out of the hardware.
i definitely felt this way with elden ring's dungeons. The rest of the game was fantastic, but the dungeons felt very bland unless they had some sort of puzzle to them (like the one where its two identical dungeons repeated on itself, that was cool), and the bosses were often ones I had fought a good few times before. I would have preferred much less dungeons like those ones, but more of the altus plateau secret passage one, since that felt really unique and had a lot to explore.
Hmmm visualy maybe but never have i though they were bland or uninteresting One of few games i had to see everything cus i wanted to I had the neeed to see it all
Elden Rings dungeons werent procedurally generated tho, its just that designers focused more on open world and smaller important sections where main bosses reside. Dungeons are just these secondary mini experiences inside grander scheme. Also I wouldnt say they ALL were that bland.
Elden Ring was definitely that game which I was glad to finish and put away after nearly 120 hours. It was right at the edge of overstaying its welcome because it was really pushing its design limits at that point and becoming repetitive. I did come back for a Seamless Co-op (mod) + Unalloyed Modpack playthrough with a friend and it was a blast.
6:00 I'm impressed with the math behind it. Procedural generation is only successful when it involves secondary systems. For example, the procedurally generated enemies in Shadow of War work out just fine. Making quests involving them would feel natural due to how the system works.
I really like use of procedural generation in strategy games like Age of empires, where you select theme of the map, but the placement of resources is always a bit random, it prevents meta gaming. In comparision in starcraft you can just learn the map and use strategy best for that map every time. I like the feeling of little bit of unknown going into the battle and real fog of war. I now it can get unbalanced a bit, but thats just like in real life and you gotta adapt to it. Best use ive seen is in little strategy game called Tooth and tail.
Random ressources spawn in RTS games can be tricky because certain players may spawn with a whole lot of ressources in proximity whilst others will struggle to build their economy
@@KalashVodka175 You always spawn with a gold patch and a stone patch over a set distance in AoE, only the location of them changes, so you still have to find them first.
For you yeah, younger people with more time do and will always exist though and what they want isn't irrelevant either. I'm glad bg3 is a 150 hour game for example, it's an amazing game. I would've been pissed if it was
"Flavourless slop" - In a year of brave attempts, there has been no finer way of describing Snorefield. That's exactly what it is; slop. I sometimes wonder what must have been going through Todd's mind in the weeks leading up to the release of this haha game. He must have been absolutely shitting himself, he must have realised it was a giant -tasteless- turkey but after all that time and all that money, he could not back out. I almost feel sorry for him until I remember that they worked on this instead of the next TES. Criminal.
Im currently playing bg3. 80 hours in. Just reached act 3. Been playing with a friend who has finished the game at least twice and already started other playthroughs and even he said hes seeing new stuff based upon my choices. Thay speaks volume to me, because ive loved this imperfect playthrough and cant wait to see how things turn out when i plau through it again. Seriously, this game is a gem
You hit the nail on the head. Procedural generation was touted as a selling point for Skyrim before its release. Radiant quest had me excited. But when the game came out and I experienced those quests, after the third one I understood they were basic fetch quests with no impact. I've been turned off from the idea ever since. Bethesda seems to keep leaning into them though so I suppose someone likes them.
@@iplyrunescape305 I'd agree with that if it wasn't for the massive amount of obsessed fans who buy it day one. If they bought it for mods, they'd wait until the modders had a chance to fix things and turn it into an actually worth buying game. The sad truth is that there are a lot of people who buy Bethesda games because they remember their glory days of stuff like Oblivion and Fallout 3, and because despite its flaws Skyrim is still a good game with only the starting of "Modern Bethesda's" lack of effort. It's the same reason people keep buying newer Nintendo games despite their decreasing quality and still praise them in spite of the flaws, or why people keep buying Halo, CoD, or Assassin's Creed. Some people refuse to give up on developers or franchises because of nostalgia for what once was.
@@Gamer3427 you are right, it's just since Skyrim the only living legacy is that it's highly moddable. No one goes like "damn I love playing through the thieve's guild and become the guild master in like one hour" or remember anything about the story. Same thing in Fallout 4. 76 required more effort to "fix" from Bethesda since you couldn't mod it. Starfield just more of the same problems, people are getting weary of this. The diehard fanboys yes, buy it reminiscing Oblivion and Fallout 3, but they drop the game pretty quickly and only come back for mods. They probably defend it still by saying "just get a mod" to "fix" problems
Procedural generation generally works best when the environment is the variable on a solid set of systems and it's not too extreme. Eg: POE / Minecraft / Deep Rock Galactic / Factorio: Maps are slightly varied, but the foundational mechanics are well conceived and consistent. Those games also have very limited focus on story and they're more about looting and group play.
Warframe does the same with its tile sets, and it works perfectly for that type of a game. Shadow Warrior 2 used a similar system as well afaik. While it was overall pretty solidly put together, the game did suffer in my opinion from having less distinct and unique level design.
It's not only boring, it creates a sense of emptiness, loneliness, inauthenticity and the list goes on. Procedural generation is decades off from being on par with meaningfully curated content.
Procedural Generation has its benefits, but there needs to be meaning there. Starfield has empty space that isn't backed by story or gameplay. If the whole game took place between a unique planet and unique moons that replaced other planets you'd have a tighter experience in theory, and the game would still be empty.
it only works in stuff like roguelikes because there the core gameplay experience that relies on doing multiple semi-short runs actually gets enhanced by having a random variance
The end of Holy Grail also works as both a cost saver, they avoid the production cost of a big battle by having the police arrest everyone beforehand but also ties into earlier in the film when the reporter is cut down, we see his widow with the police.
The primary reason us old timers love baldurs gate 3 is that we played bg1 and bg2 20 years ago and creamed our pants even knowing it will take us years to beat it
@@carsonspears8568I’m 45 years old and never played a crpg before. I’m now 50+ hours in bg3 and I’m loving every minute of it. It’s opened up a crpg rabbit hole for me now, lol.
You hit the nail on the head around the 5:53 mark. I was excited for Starfield and played it a lot when it first came out(and still plan on finishing it, or the main story quests at least), but the more I played, the more I saw how randomly visiting planets and systems were basically pointless. Sure you could set up outposts and farm resources, or you could do what I did which was loot all the weapons and armor off enemies, vendor them wherever I could, and then just... buy resources from a vendor because it was just easier and you didn't have to go out of your way for it. Completely takes the point of resource farming out of the equation. The weapon and armor system also feels like a half-baked version of Fallout 4 and 76's versions. Melee and unarmed combat is pretty much pointless because you cannot mod any melee weapon, or mod any spacesuit to allow for more unarmed punching damage. There's no Power Fist equivalent either. The guns feel fine to shoot though. I've seen the phrase "a mile wide and as deep as a puddle" for Starfield discussion and it's 100% true. The story is passable, and I've been enjoying the faction quests and other sidequests I stumble upon(Operation Starseed in the Charybdis system for example), but once I finish the story, I'll probably shelf Starfield and go back to Skyrim or Fallout 4 if I want an open world Bethesda experience. Also, who consciously decides not to make local city maps? It's easier to navigate them once you become familiar with them but it's still baffling that there's no proper city maps. :
Have to disagree on the side missions. Most are meaningless fetch quests. You get better side quests in MMOs... The faction quests are the only missions worth playing. The main plot has exactly two missions that aren't fetch quests. I am baffled that the UC Vanguard plot is not main missions. What airhead at Bethesda thought that new game plus makes for a great main plot?!
@@haariger_wookie5646 That's fair about the side quests. A few stick out for me even if the objectives are basically fetch quests, like Operation Starseed. I agree with the rest though. Funny enough, I picked up No Man's Sky on a sale this past weekend so I'll be trying that out once I finish Starfield's story lol
Thing is bethesda sucks at most things. They suck at story telling, character writing, making interesting gameplay... Their main strength is world design. It's the one thing they've been consistantly getting better at, even with FO76, and with starfield they just threw that out of the window. Even the handcrafted places are impacted because they needed to tone down the designer's creativity in order for it to fit seamlessly with the procedurally generated parts.
If ressource farming was mandatory people would be saying it's "forced busywork"... There is way more weapons variety in Starfield than Fallout 4... The quest of Starfield are miles better than in previous Bethesda games, the writing and worldbuilding in general is very good, this game is pretty deep. City maps would just make it so the player spend more time in the menu, it's more fun to learn the city yourself or to use the in game indications
@@haariger_wookie5646 That's just false, all of the side quest have some sort of narrative, a fetch quest would be a quest without any narrative. It's like you never played RPG, in most RPG the side quest will involve going to pick up an item, clear a location or kill someone, but like I said what really matter is the narrative of the quest.
When I was a kid, my favorite show was ER. My all-time favorite episode of ER didn't take place in a hospital. The entire episode was shot in a small classroom with only 5 of the massive ensemble of more than 15 actors. It was great. The drama was intense, the emotion was high, the dialogue was biting, all top-notch. The intimate setting made that episode what it was.
I feel you. I'm still in act 1 of baldurs gate because i play it slowly, take my time, try to see as much as possible and just enjoy what was given to me.
Its also like 125 gigs to download and takes up 140 GB of hard drive space. I have no idea how they managed to make it a bigger size than Red Dead 2 but they did. Devs should really start making 2k/4k textures an optional download (Like Diablo 4 has) cuz these file sizes are getting insane.
Honestly what I am really impressed by is the procedural generation in remnant 2. They wanted to make it much more focused around center points so that the game has a focused, enjoyable plot but has some variety when you replay the game which makes replays much more enjoyable
even there, the dungeons have only a few different layouts, most of the "randomness" is which dungeons you're going to have spawn in your playthrough. Its very good because most of the game is intentionally designed, there's just a bit of shuffling to keep things fresh.
Remnant 2 does it well because it's hand crafted tiles, that are then shuffled around within parameters. It's not truly procedurally generated like a game like No Mans Sky. After you play that game for a hundred hours or so, and constantly secret hunt. You start noticing patterns in the game and can hunt areas down based on landmarks.
Remnant is the only game in which I can't wait to see how they keep mastering their way of using Procedural Generation with subsequent releases. Absolutely a banger of this year and to think we haven't even gotten to the first expansion but yet have such a kickass base game to play.
It also helps that Remnant's fundamental gameplay loop is well-tuned and extremely entertaining - despite the procedural generation, there's a lot of GAME there, and as you say the procedural elements they include are curated to augment that rather than strictly being designed for scope. The problem becomes when a game (like Starfield) throws together a bunch of mechanisms and crosses their fingers that a game will somehow appear.
85 hours in Starfield and it feels like a drag. I had 120 hours in Baldur's Gate 3 and it felt like it was only 30 hours. Baldur's Gate 3 is out of this world amazing.
@@henryjohnson-ville3834because deep inside down he believes that something interesting is gonna happen. Also since you invested so many hours in this game you want to finish it for sure.
@@revbenSame happened to me with Skyrim. My bare maximum is somewhat around 50+ lvls, after that I got so bored that just deleted the character. The game itself is just a clicker without any gameplay itself, without story, graphics or mechanics, but for some strange reason I was thinking that this is going to change somehow. That was 2 years ago. Well, the main logic behind this torture is just an idea that you will see new enemies, new locations, new something. I felt sorry for spending hours into this, like really sorry. But sooner or later I had to admit that's not going to happen, it's just empty and never ment to be filled with anything. So I think that's the reason why people play "drag" games. Coz they "invested" so many hours that they can't admit it was for nothing.
The first games I got into were traditional roguelikes, and my opinion based on that experience is that procedural generation only works if you fulfill a few requirements: - Handcrafted (or very constrained use of generation) setpieces. - Mechanics focused on mastery. - Procedural Generation that can facilitate those mechanics and setpieces. Brogue has some of the coolest level generation that I've ever seen.
Yeah, my dad's been hooked on Spelunky for years - he's not really a speed runner type but the depth of mechanics mastery and understanding combined with unique situations where the generation puts problems next to each other makes for his crack (I'm only not a fan because I can't platform to save my life, especially in 2D) I think the game's strategy of having handcrafted 'rooms' that get smudged, shifted, rearranged, and randomly altered gives a good balance between the maps feeling unique and the maps feeling intentional
I immediately thought of Brogue as well, also because the dev had a talk about proc gen in Brogue (it's on youtube), where he talked about developing proc gen being a skill as well as designing levels by hand is. The proc gen behind Brogue is a craft, the reason behind the huge(lava)lakes and chasms for example is because being able to see a bit of what you'll go to later on makes exploring fun. Seeing rooms on the other side of the lake is similar in that sense to seeing a tower in the distance in an open world game.
@@madeliner1682 You and your dad might also like "Unexplored" (the first one) then! From what I heard it has a good proc gen system made to make exploring fun (inspired by the way the dev's kid liked to run around the room when playing), and it has a similar focus on problemsolving and understanding mechanics. Also, it's top down, not a platformer.
I'm 37, full time job, married, make a decent living. Just finished the Mass Effect trilogy again, and enjoyed every minute of it. Started Andromeda, and in 18h decided it's not worth my time. I'll gladly dump 1000 hours into something like Monster Hunter World. I refuse to waste time slogging through a mediocre game like Andromeda, because "mid" is not worth my limited time.
A couple months before Starfield came out I was talking with a friend about it. We'd both loved Skyrim and the Bethesda formula when we were in high school, but we both have Playstations so he was bummed about not being able to play Starfield. I said I wasn't really hyped for it because I expected it to be repetitive, way too big for my schedule, and full of filler content, even more so than Skyrim was (and I couldn't play that either if it came out today). He didn't believe me and thought I'm just trying to convince myself I'm fine with not being able to play it. Everything I've seen of the game just convinced me of that first impression... Your point about non-gamers being impressed by buzzwords also hits a note. You see it with VR and the metaverse, with techbros pushing how immersive web3.0 will be because it's big and realistic and you have to actually go from place to place physically in the digital world etc. and like, those are inconveniences in the real world that I'd be happy to do away with in the digital one... It's what made RDR2 not fun for me, when instead of a store menu UI they make you physically walk to the product you want to look at. Or like early smartphone UI conventions going for realism and retaining the inconvenience of what they're simulating, before realizing the potential of digital space to be easier to interact with. Anyway sorry for the rant, I work in tech and am surrounded by non-gamer tech enthusiasts who don't know we've actually learned many lessons that go against how they imagine web3.0.
Gothic 1 is a 2001 RPG where creators had to come up with an idea onto why the location is small and limited. They made it into a magical sphere prison and suddenly technical limitation became a narrative feature
It also didn’t have loading screens. You could go anywhere, anytime, every NPC had a routine. The only loading screens happen when moving to the next story chapter, at which point parts of the world change.
@@ArDeeMee , well, in Gothic I there were a few separate locations behind the loading screens... Old and New mines, a couple of temples. Same in the Gothic II. Still, the Colony and Khorinis was (is) a living-breathing experience you can immerse yourself into. One of my favorite RPG's! Among Gothic series, I think part III has a truly loadscreen-free world. But I love parts I and II more.
@@leh2012 Yeah, the few loading screens happened every time the world changed permanently, e.g. new chapter. In G2 it happened at those points, and if you went to the colony, aka an entirely different map. It’s the first game I ever saw where each and every NPC has a daily life. It was sooo good.
@@ArDeeMee , I was amazed by those game too... A good momories! But I think it can surprise even today those who missed it beack in the day (or those who are younger than the game). If they manage to run it on their modern PC's that is. Also, there are lots of mods available for the game some of which are basically a new games...
I think the procedural generation in Minecraft was and still is impressive because its the main soul of the game. Probably the only game I can think of where it is essential and makes the game what it is.
For something that's essentially the opposite of Minecraft, look at .kkreiger from 2004. It's a single-level FPS with graphics on par with Doom 3 that fits into 96KB. The level layout isn't procedural, but everything else is and is built from instructions at load-time. Not that interesting as a game, but definitely a technical marvel.
Josh's point about limitations breeding creativity is actually a big part of why Artix games are so great (some of them, I mean). One of the developers of Dragonfable was talking about this exact matter only a few days ago in the Falcon's Nest discord server, they're forced to get creative with character models due to the limited number of animators they have. (PS We at FN are eagerly waiting for the DF video, Josh!
Doctor Who's penultimate season 4 episode "Midnight" was written because they had to replace an episode that wouldn't be done on time, so all they had was a bus set. And they made one of the best episodes ever. With no props, little CGI, and no money. All excellent writing and using limits.
You nailed it once again Josh. Procedural generation is good for certain games but not all. BG3 is an example of a long RPG I want to make time to play due to how compelling it is and how when you explore you find quality side content and interactions. Starfield has no appeal to me. I do enjoy No Man's Sky but it's a game I play on and off when I want to wander and chill out.
I would like to say that an example of where procedural generated content works in an RPG is Wildermyth. Each world in a campaign is made from scratch, your characters can be as customized or random as you'd like, and beyond the overarching campaign storyline, everything that happens in between is randomly determined as you go. Obviously it's not 100% random, but each campaign promises to be a unique experience.
What I hated the most in Starfield is not the emptiness but the missions are extremely linear. Even when the game gives me different dialogs I realize no matter which dialog I chose the result was exactly the same. Ex: we had to deal for an artifact and there was a dialog option to attack the man I was completely deny to attack him. No matter which option I chose the end result was exactly the same. I only had the illusion of choice. There is the pirate mission where my partner wanted me to side with him to take out the leader. In the end it made no difference, the end result were exactly the same as if it didn't happen.
When everything in the game is up in the air, how can the missions be anything else but linear? Think about it - to implement a story mission, you have to know the locale, the layout, everything about the area. When it's all a diceroll, you cannot reliably account for anything that the dice will affect - so you can only write absolute barebones generalities and HOPE the dice roll some cool stuff. Sometimes they do, sometimes they don't - but most of the time, the dice roll their mathematical average.
The artifact quest really killed the game for me, I know Bethesda well enough and I didn't expect much freedom of choices, but that quest showed how thoughtless the quest structure is. I talked to the guards and got some hint for "hidden path", then I managed to persuade the collector to show me the artifact, I thought I could pretend to leave first and sneak back since I knew the hidden path, then I realized that the collector just stay at the vault and never leave, I back stabbed him and found out that he cannot be killed instantly and will always trigger the dialog of "take the artifact and get the bounty" or "no witness but get the bounty anyway". I loaded the save and pick pocketed him to get the key, then sneaked to the artifact, "bounty added". I reloaded again and killed everyone on the ship and took the artifact, "bounty added". I know the bounty was meant to trigger the Crimson Fleet quest, but the choices make no sense, why the quest ask you to "steal from the thief" and "do it peacefully if possible", and why was I getting the bounty when there were no witness or even no one alive.
@@thesunthrone no, lack of choice even appears at the end of unique side missions, which are irrelevant to the main story. The OP has a point. I don't mind most of the flaws, but that's just lazy. Far too many characters are immortal - perhaps they need to be used for DLC, but even so, that's lazy/cheap too. Anything that's worse than Skyrim/Fallout in Starfield is bizarre. I guess they started off too ambitious, broke the engine, so had to remove stuff - a bit like GTAIV after GTASA, but not as extreme (and not a different engine). Most of Starfield is typical Bethesda - people were expecting a brand new game? Naive - it was always going to just be a Fallout reskin. It's a good game, Brent - if you like Fallout/Skyrim, albeit nowhere near Skyrim's level, although combat's more fun. Typically addictive and atmospheric, but typically technically poor too (compared to GTAV, RDR2, and Halo Infinite, which should be expected).
@@goatlps Don't mistake me for excusing Starfield - I'm just saying what procedural content means. When you roll two six sided dice, 50% of the rolls will add up to an average of seven, that's just statistics. But what good are quests that can work only 50% of the time? The procedural design necessitates that quests IGNORE the procedural design entirely for them to even work - and because of that they'll just be the most blandest, most boring fetch quests imaginable, because nothing else can even work.
@@thesunthrone makes no sense. You can create procedural environments and interiors and still have good quests. They just didnt bother making good quests. Except a few ones, like the one in a lab where there's a dude stuck in some kind of other dimension and you have to get through portals to naviguate between both realities. Environements are procedurals, quest is interesting enough.
Procedural Mirrors Edge would be awesome! One of the reasons I enjoy speedrunning missions in Warframe. Great movement/parkour through constantly changing levels.
Procedural generations works in Warframe because of the gameplay loop. But, DE didn't just throw some random prompt. They created each object and room by hand according to the theme of the planet or level and then let the game decided which room connect to which. Yes, that occasionally would create a nonsensical level design like a 2 or rarely 3 or in very rare case, 4 identical room created one after another and weird dead ends that logically shouldn't built like that in a space ship. But we just accept it is as what it is because tbf we don't gaze the scenery. We grind til morning for materials.
The Deadpool fight scene situation is so perfect because it's completely within character for him. They managed to make a limitation work with an established character. That's great.
Thats one of the things I really love about Breath of the Wild and Tears of the Kingdom. The map is just packed full of little curated details. It's amazing.
Yup, and despite being on the Switch’s archaic AF hardware, the world is more DETAILED than 95% of open-world games (and I DON’T MEAN in terms of visual fidelity because it’s easier to make something look nice and flashy as opposed to making sure the “world simulation elements” can respond to any number of things you can do).
For BOTW I can agree. Totk though, even though the sky islands are fairly good, the underground is pretty awful and barren. This is made especially obvious when you start lighting it all up with the light roots
@@fuzzydude64 Well, you can KINDA say the same thing about the shrines in BOTW where some had FAIRLY BIG SPACES within them but not a whole lot to interact with outside of whatever elements pertaining to the puzzle you needed to solve there.
Level 1: Hay Bale Level 2: Hay Bale + Tractor Level 3: Hay Bale + Tractor + Building Level 4: Tractor Level 5: Tractor + Building Level 6: Tractor + Building + Hay Bale Level 7: Building ...😂
To be fair, the scale of games used to be impressive. Wow was mind blowing because of how big the world was and how much time you could spend on exploring it through different factions/classes. Now, we are more conscious of the padding and the repetitivity of these games.
Hi-Fi Rush is a perfect example of this. Hi-Fi Rush will take about 8 or so hours to complete, but the amazing gameplay, music, and visuals will make you want to play more. In my case, I went and spent 82 hours getting all the achievements in Hi-Fi Rush. I wouldn't have done that if Hi-Fi Rush wasn't as good as it was. I chose to play Hi-Fi Rush for 82 hours, the game didn't expect me to slog through 82 hours if subpar gameplay and visuals to get to the end of the game.
As for budget limitations; I think it was DS9 where they had to scrimp and scrape the first season and one of the best, yet cheapest episodes from that was Duet. Brilliantly acted and written with no grand battles or shoot outs or anything not even phaser or disruptor fire. Just good solid dialogue.
You're not wrong about the game-length aspect. BG3 for example by the time I got to Act 3 I was running to the finish line. I'd experienced everything gameplay wise it had to offer, most companion stories had wrapped up - it was just the story left so I didn't even want to explore any side content just reach the end already.
Agreed completely. The only game with a narrative that I've played where I enjoyed the procedural generation was XCOM 2/WotC. Its absolutely amazing because it forces you to adapt your tactics, but they made sure geometry doesn't break any more than it did in XCOM:EU, and it is a strategy game. I can't say the game would have been better with set maps because it would come across the issue of every mission being the exact same, and I honestly wish more tactical games were similar.
I think one of the other big problems is that the Procedural Generation engines people make tend to be way too simple. Like you said, start in a forest, there's an object you'd see on a farm, there's another one you see on a farm. The engine didn't make significantly different ones.
Too simple' does not reflect the essence of the matter - AI will not write a good book for humans. It can only make monsters out of pieces 'of other people's work, but he won't tell the story because he doesn't feel it or understand it. Moments full of emotions in games (or music or movies) are a symphony of emotions and means of expression (dialogues/words, sound and images) - AI is not able to recreate it.
@@mwmv921 yeah but not every moment in a game needs to be a big deal dramatic climax - there a lot of value in allowing less plot relevant experiences that let the player understand the themes and mechanics of the game. Those sections can be procedurally generated, if you do it right. But to do so, you need to take the time to design the hell out of those systems, making sure it produces sufficiently different but related experiences every time. An entire game cannot be made by a computer, but I think there are a lot of segments that can be, if the devs took the time to do it right
Early No Man's Sky was a lot like that. Thankfully, I bought it used a year after release and not at release, but after about 30 hours in the grind became boring. I still revisit it when there's updates, but exploration doesn't have any appeal to me in NMS anymore. Contrast with Minecraft, where exploring procedural caves still feels interesting because of the challenge in it (especially in early game)
Imagine a space exploration game with the same budget, but instead its based in one star system and just has four or five planets, each of which has a rich variety of biomes, factions, locations, characters etc on it to explore, plus some orbiting stations and asteroids and stuff. That would be so much more immersive to me than any million procedural planets in this or no mans sky
I mean, it could be two or three systems even, with 20 or so curated planets with a hypergate or some such in between. Could still have the scale but more of the quality that makes a good game.
Starfield has a rich variety of biomes, factions, locations and character to explore. Starfield also has orbiting stations and asteroids. All of the content in Starfield is handcrafted, the procedural generation is only for the landscape.
It's good to see that Josh's Mic has a procedurally generated waistcoat.
I didn't even notice.
Lol!
Neither did I, but I see it now! 😅 It's the one he uses when it's too hot to wear his usual vest. Lol
lol
Pretty sure that's bespoke, sir.
"Wide as an ocean, deep as a puddle."
Is my favorite quote for these kind of games
No Man's Sky was the same for me.
@@kane-111 At least in NMS you can _feel_ the efforts even if you don't like it.
And don't forget that on every N step you have to _load_ more puddle. Also remember that it's not just about the width and depth, but also the _variety._ Color, flavour, atmosphere, etc.
TB was quite right when he said that quote for Skyrim. He would probably say the same for Starfield
well you're so clever.
I respect the hell out of this man for saying he doesn't play games he really wants to enjoy on stream. It's like people forget that streaming is a job for them.
Im mixed with this statement. On one hand, you got to play the game you love however you like. On the other, you practically admit that you dont enjoy the game that you actually stream.
@@RhazOfRheoshe didn't say that. Your assuming that from his words. I think he enjoys games in general then they're are games personal to him he wants that time to himself not sharing it with the stream.
Isn’t him saying this in relation to the fact that BG3 is a long game, he wants to finish it, and if he streams it he plays at half efficiency (due to like chat interaction and such)? It’s not that he can’t enjoy it, it’s just the game will take longer.
@@SeriouslyNotMichaelJordan he started playing Dragon Age Origins on stream a while back, but decided to finish it off stream so he could really get into it and enjoy the story without worrying about interacting with chat.
It's not really to do with it being a job, it's that it's harder to be truly immersed while you're streaming. This works like this even if you're streaming it to a single friend.
"Some people might enjoy the extremity of the flavourless slop"
"We live in the UK, Josh... come on"
Your chat is a treasure.
*Taybarns has entered the chat*
lol I was going to comment on the exact same moment
says the american
@@macbethhm Hey, we have artery-cloggingly, diabetes-inducingly delicious slop!
Starfield's ultimate problem is that it's not just a lot of nothing, but the fact there is stagnancy and corporate safeness. Like there are essential NPCs in places despite the entire game being about the multiverse, there aren't any mechs despite there being independent space, you can't edit the interior of your ship despite ships bing the main advertisement of the game.
Starfield's ultimate problem is that it's a Bethesda title that they "worked" on. Peel back the new coats of paint and wallpaper and it's literally a Skyrim clone with less content. The problem is the whole video game industry, mainly AAA stuff, has been going fully corporate for years now, and we're pretty much at the end game for that. I wouldn't expect anything traditionally good from them anymore.
@@ericlamb4501 all Bethesda games are just more bare bones versions of Morrowind
@@schwarz8614are they still reusing the werewolves animations from Morrowind ? It was still there in F76 😂
@@CatroiOz Just guessing, but are we talking about the swipe-swipe claw attack animation?
I saw a great example of a random quest:
Go find our missing scientist! He's 400 m away and you have to walk there. Nothing happens or gets found on the way. Arrive at the cave; dude is just standing there not in danger. No escort back; he just walks back on his own. Pass Go, collect $200.
Skyrim's random quests pissed me off because you get the same pittance for killing a Dragon or a Bandit Chief at level 1 as at level 50, but at least there was some excitement, found a new shout word, found a new cave on the way, got some good treasure, whatever. Starfield seems even lazier than Skyrim.
Fully procedurally generated experiences in RPGs often feel like your in the same house just with the furniture moved around.
Im feeling like this in almost every Bethesda game.
They are only building a library of asset blocks and reuse them over and over again.
Every house, every cave, every dungeon, sewer, cabin, subway looks and feels the same from the inside.
And so do the 5 outposts and caves you find in starfield over an dover again.
In Starfield they didn't even bother moving the furniture. Literally. You can find the same notes and named corpses in copypasted facilities on different planets.
@@argonianmate3191ewww
@@leinadreign3510 At least in most of their previous games they touched them up by hand or often had at least something interesting in it/going on or a quest related to it. Starfield is way worse because of the procedural generation you don't even have that much and random exploration feels a lot emptier because if you land somwhere random there isn't even the chance to stumble upon an interesting quest, item or location because the only hand made stuff is in the pre marked locations and quest locations. At least so far I haven't found anything in a random location I couldn't have found while doing the quests.
Hey , over at is having trouble with some and they've asked us to deal with it. Go do that, then come back to me for your reward.
reward = levelled variant of some semi-random generic weapon
There's another story about Monty Python's The Holy Grail that I feel fits in well here:
While filming one of the scenes in the swamp, production was slowed by a faulty fog machine. After a long while, a frustrated John Cleese yelled to the crew, "Is the fog funny?"
The moral: if an element to your work of art doesn't exist to serve your ultimate point, then why worry about keeping it at all?
I recently saw a video on the making of Monty Python's The Holy Grail and the biggest reason the ending scene happened is that they just couldn't afford to have a large battle scene which the "cop out" was the solution to.
Was every element in that film design to be funny? The jokes wouldn't have landed without the context of the setting being accurate enough for the subversion to happen in the first place.
@@daruddocklike 90% of it was yeah.
Nobody would probably see this comment, but I definitely agree with "limits breed creativity". I am a comic writer, and I used to have this issue where I would just fill the pages with useless fluff, and I would feel my stories walking in circles unable to fully progress, so I told myself "ok, you will now only have four pages per chapter. Do what you will, but you must have a complete story arc in just 4 pages". And I started to think about what I actually wrote, because I didn't have the luxury of space. Every panel, every line of dialogue needed to mean something, because if it didn't, it would just take space from something that did. That's why I say "Yes" one hundred times to limits. If you don't have natural limits to your work - create them. And that's what a lot of modern games don't understand, and why they start to wander. Why the feel watered down
A pity the TPB killed the notion of self-contained issues for DC and Marvel.
I see it and I agree. I dabble in writing stories in my spare time, some fan fiction, some original stuff (mostly Fantasy) and some D&D campaigns. Core to all my work is always to look at creating the boundaries, the walls and ceilings for whatever powers will be involved in my story. In Fan Fiction I can look to whatever I am writing from. It may not always be stated explicitely, but looking at the application of powers or other parts of the world can often inform you. In paticular, the lack of something can often tell you just as much as its presence.
Hypothetical example. You have a world populated by people who all have some kind of elemental power. Everyone has one power and only one power. Logically the fact that all people only have one power would probably mean that offspring of two people with different powers only inherit one of these powers and is not a hybrid of some kind as, assuming some amount of mingling between those with powers, if hybrids were a thing, one or more would have popped up by now.
Now this is more a in story limit rather than a creation limit, but I feel they are just as important cause if you dont set those limits on the world then you might be tempted to just bend the rules if you write yourself into a corner...and suddenly the Fantastical but logical world you created starts to crumble.
I haven't seen this comment, but agree as well
You don't need to start your comment with the assumption that nobody is going to read it. Just write your comment, and chances are that people see it.
@@janelantestaverde2018 hahs, I guess, I really didn't have too 😅 though that is my usual outcome, so I just assumed 😂
Deep Rock Galactic is based completely on procedually generated Levels and its friggin amazing. Really feels like a whole planet full of caves, tunnels and glyphid hives.
That's because each room is actually designed mostly bespoke. They just string them together and procedurally generate some geometry and interactives inside those rooms. The actual gameplay loop is pretty tightly curated.
Deep Rock is almost entirely driven by gameplay and teamwork, means that the short falls of procedural generation with story/characters/plotlines doesn't apply at all thankfully
Sounds like I have to give this game another try after like 3 years or so, because gameplay was just as repetitive and boring as the generated worlds back then when it was rather new.
There it is, someone mentioned the best game.
Rock and Stone.
Yes, and is a fantastic example of the gameplay must also be good for procedural generated worlds to work.
There are some absolute fantastic examples of it working really well, such as roguelikes etc.. but when someone uses it as their unique selling point, it's a massive red flag.
Deep rock galactic has so much charm and other features that most players wouldn't even know it's procedurally generated because everything else is pretty good.
Deep rock galactic was one of the few games that procedural generation got right. Every cave felt and looked different, then you play for 50 hours and see that what you actually get are a random selection of repeating rooms, spaces, corridors and connectors. In a game about exploring and spelunking maps knowing your way around and where all the resources are removes half of the game
The thing with DRG though, is you'll never have the same experience twice. You might run into fundamentally the same rooms, but the moment you remember that you're in a cave system, you kinda have to remind yourself there's only so many ways for caves to form, and that helps cushion the idea of repetitive rooms, especially when your Engineer this round is trying to dig a hole straight to the bottom of the map and isn't helping with anything, but he digs a massive hole and you are compelled to go check it out, because fuck it, what's he actually doing? At that point the mission becomes secondary because players are using the sandbox that's available, and you're just curious as to why there's nothing quite like elsewhere, because in that game, the players you get along with you are what determine whether or not you're going to have a fun game, and you virtually never run into a hostile player who wants to ruin things because you're in a game that encourages positivity.
@@nuclearsimian3281those moments are the best thing about drg, the moments when the mission goes out the window in favor of yelling "we're rich!" Or pondering the orb. It's such a great design that it's addicting.
Rock and Stone!
I think that games which take heavy advantage of three-dimensional content can benefit the most from procedural generation. Deep Rock Galactic does this very well, but Minecraft also does very well with its numerous biomes and things like cave systems to explore. With a third dimension added to procedural generation, it becomes more complex to render and piece together properly, but the result often can be spectacular. I would like to see more games taking advantage of that kind of thing. Starfield with some procedurally generated caves that are worthwhile to explore definitely would have been much better, at least.
I love DRG but the caves absolutely do not look or feel different. They are made different by the chaotic and dynamic gameplay usually by the end of a run, but when you first launch in, other than mission type, very little about the levels themselves is different
Most roguelikes in general are a good example of procedural generation...and that's why I love almost all of them even with the punishing difficulty that comes with some of them
Starfield is a bad implementation of procedural generation...which reflects the title of this clip which the game is filled with (literally) a lot of "nothing"
I immediately thought of Subnautica when Josh was talking about this. Initially (in alpha/beta I think), it was going to be procedurally generated.
However, when they added in the story, they chose to cut the procedural generation in favour of more curated, thought-out world design, something that worked out greatly in the game's favour (in my opinion, at least).
Many locations of key items are still random
@@Aggrofoolloosely random, anyways. Only resource nodes like rocks are random. Shipwrecks and data module caches are fixed positions.
On one hand it gives an amazing first gameplay experience especially when you talk to other people and they have similar experiences as you such as the reaper leviathan by aurora, or the "multiple leviathan class creatures ahead, are you sure you want to continue".
On the other hand, the game has 0 replayability and once you complete it, you can't help but feel like you want to play more but subnautica doesn't scratch that itch anymore. So you feel kinda empty for a time, try find similar games but nothing comes close to subnautica as it genuinely is the top of that game genre.
So honestly it's just a difference of opinion and perspective. I don't think one type of game beats the other, but it depends on the individual and what they desire in the moment. I personally like a balance, I play a lot of grindy games (not always necessarily procedurally generated but not lore heavy either) and then some lore heavy games. So there is a need for both types of games to exist.
@MrTalithan strictly speaking, the fauna are procedurally generated by and large. There are rules, obviously, but it helps keep the world moving without blowing out your processor.
@@MyNameIsSalo "the game has 0 replayabilitiy" and yet most people I have talked with about the game have 2-4 playthroughs. To me it seems that on average it takes the player 2 playthroughs to really get familiar with the world, as you can easily miss a lot of stuff in the game world during your first playthrough and the second one becomes much more about just the exploration of the world since you already know what you have to do with the plot so you feel like you can take more time just looking around.
And then there's the people who start new playthroughs just to try and build their first base on whatever biome so that they can look at leviathans through the windows.
One thing that shocks me in Starfield is if you're in a major town or city you can shoot a gun one inch from an NPC's head and get zero reaction, like GTA 3 had that figured out in 2001...
Even worse, Bethesda themselves figured that out before too. They are regressing, which makes it all worse.
@@Pangora2They we’re too busy jacking off on their procedurally generated soup of a world and had no time to consider actual immersive role playing elements.
@@Pangora2 it makes me reeeally worried for ES6… Elder Scrolls is my favorite game series and the only reason I still follow Bethesda’s projects..
You'd think that these major game studios would have learned a thing or two from the awful state that CP2077 came out and all the backlash.
@@junojuno3745 Before Cyberpunk there was Anthem, Fallout 76, Watch Dogs, The Division.. etc… games that come out and are stable and polished at release are becoming the rare exception sadly
I was genuinely impressed by procedural generation back in Elder Scrolls Daggerfall. The idea that you were walking around on a landmass the size of a real country was so cool to teenage me back then. Of course, we always fast travelled and in the end it wasn't anything more than a gimmick.
There’s a reason we invented cars. ^^
Bethesda has relied on gimmicks to sell their games in the past.
That was a 90s game though. We could do a lot better today.
It had potential to be so much more, but sadly they abandoned that and when they finally went back to heavy reliance on that technology, they do nothing to add substance with it. And personally I still prefer the way that world feels more real even while just using that fast travel system. It manages to facilitate immersion with it's systems rather than a dull feeling of emptiness
@@ArDeeMee if only cars exist in Starfield... or bicycles..... or at least horses.... Well, they have space cowboys 😀😀
Being an IT guy for over a decade, and having been well into my career by the time "procedural generation" became a buzzword, has inured me to ever being impressed when a game uses that term in its marketing.
However, to answer your question Josh, I will say that I have finished exceptionally few open-world map-marker driven RPGs, and the reason for that is because when I start a new one, I often feel that rush of excitement and anticipation: "Oh wow, this world is HUGE, there'll be so much to explore and discover and do, I can't wait!".....And then after scaling the fifth tower to uncover all the little markers on a new part of the map, and seeing that the number of "go here, collect/kill/mark X" icons dramatically overwhelms the number of actual bespoke side quest markers, I eventually come to the same realization every time--it's all padding to turn a 20 hour game into an 80 hour game, and not only is it boring and repetitive, but in most games all those little "to-do's" don't add anything to the gameplay, story, world, or characters I'm trying to enjoy in the game. If anything, they dilute it to a terrible degree.
The only difference with procedural generation is that, as a developer, I know the nuts and bolts behind it, so I know to expect that feeling from the get-go, and consequently avoid any RPGs or adventure games with that as an advertised feature.
I would say that open world games really rely on gameplay and the world itself. Games like The Witcher 3 (I know that people complain about the combat, but that game has a beautiful world and the story is fantastic) and even Horizon Zero Dawn do this very well. They are not procedurally generated (as far as I know at least) but they are great open world rpg games. No Man's Sky has improved a lot but the procedural generation of it is hardly even the part of the game that people really care about. Yes you do get a bunch of interesting planets out of it, but for the most part the only difference is some color filter at some point
pedo clown say what?
I think procedural generation has it's place, and can be impressive in right spots. But can't work everywhere. I think for an RPG, it's generally a bad idea. Unless in specific circumstances or sections. Like with ARPG's. But I would add that procedural generation can work wonders in certain strategy games. Civilization, Age of Empires, Heroes of Might and Magic (random maps), Stellaris etc. work very well with "endless gameplay" and procedural generation. And they can easily be 1000 hour games, where all gameplay feels tight.
But ofc - those are all games where - a single "session" is not 1000 hours. It's only 20 min to couple of hours long. Some, like Stellaris can have longer sessions, but it's still not more than a dozen hours per. Similar with Roguelites. Procedural generation makes sure every try or every session feels a bit different.
But a classic RPG, like Skyrim or Baldur's Gate - you aren't resetting your game every hour. You play a single session that is 20, 50 or 100 hours long. And if you stretch this too much, by adding procedural generation, so that experience lasts 1000+ hours, it will feel "empty and pointless".
I would add that I think procedural generation could work in those games, by only applying it for a few sections - like dungeons in Skyrim for example. If they were procedurally generated, it wouldn't increase the time it takes you to finish the game. But it would make every time you restart the game, feel a bit different. Thus giving a bit more re-playability.
@Voidstroyer Exactly. And to be clear, I have no problem with devs using procedural generation for stuff like, say, building the base geography for the world map, or even populating that map with certain features like forests, etc. But the actual locations and wildlife and characters and quests, etc, down to the prop models, should be placed in a bespoke fashion by a designer with intentionality and purpose, otherwise players will feel the meaninglessness of those elements when they play the game, even if they don't know enough to identify why those things are clashing with their suspension of disbelief.
@Member_zero Yup! That's why I specified adventure games and RPGs. Obviously procedural generation is a fantastic tool for stuff like roguelikes/lites, as that's a genre that would probably not exist as it does without it.
"A 1000 hour game is useless if you don't want to plat it for 1000 hours" That is such an underrated quote
His question about not wanting to play a game if its 500 hours because you don't have the time. It doesn't apply to myself. I do have the time. My problem is that I can never stick with a singleplayer game for longer than maybe 20h. Unless its a multiplayer game where I keep getting engaged by other players, I just lose interest.
@@DeltaNovumsounds like a you problem
@@amadeux5471 that's a very astute observation. It might have something to do with my adhd. And I'm not the only one.
the obvious is underrated?
A good game naturally makes you addicted in the most fun ways
I get what Josh means about how Nothing can be impressive in certain circumstances. The Glowing Sea in Fallout 4 was like.... 90% just featureless wasteland, but man did it nail the atmosphere of an ungodly irradiated hellhole so hostile to human life that it may as well have been an alien planet. Even if it wasn't actually dangerous to an even moderately prepared player - and let's face it, it wasn't - it was just so damn cool.
bethesda fan boys being impressed by the bare minimum
in my opinion the area before the Glowing Sea, around the south of the map was much much better vibe wise and everything compared to the sea itself.
@@steaketc2476the only place Bethesda isnt doing just the minimum is the map design (pre starfield, even FO76s map is fun to explore most of the time)
The Glowing Sea was so disappointing. I remember feeling unfulfilled after completing my exploration of it.
Playing Baldurs Gate 3 for 10 hours felt like 1 hour.
Playing Starfield for 1 hour felt like 10 hours.
😂😂😂 True 🎉
I wish he will stream Kingdom come deliverance. After years of listening I am quite sure he will enjoy it
Honestly it fucking did, I specifically remember the weekend I first played it, I woke up around like 6pm (nightshift worker) and played through an entire questline, made some money and eventually ended up with a ship at the stats I wanted and finally took a break, only to notice the sun was still fucking up outside. It had been like three fucking hours and I felt like I had been playing for like half the night already.
If fun and enjoyable games suck away my free time, then that should have been the biggest red flag that I was playing a very boring, checklist of a game that isn't engaging me at all. I was just engrossed in the novelty and hype of a new game, a space game at that, which have even more sway over me.
I just walking around in one small place for Baldur's Gate 3, barely progressing to a quest, and it takes me somehow 6 hours, and it's still puzzles me why. Going on to one Skyrim dungeon for an worthless RNG loots, and the dungeon takes 30 minute to slog through.
Doesn't want to bother buying or playing Starfield. Because afraid that the crappy old Bethesda formula is still there.
you have terrible taste then
It's a funny point you make about procedural generation.
I once worked on an RPG project where, as part of the setting's lore, there was a massive forbidden zone to the north--a blasted hellscape chock full of dangerous creatures, but nothing else.
There was no xp to be had. No quests. No loot. Just an endless expanse of procedurally generated wastes and deadly monsters.
In our case, we actually used procedural generation to discourage exploration. The entire zone was completely impossible to map out and completely pointless to explore.
Yet, there would be people who explore this area nonetheless.
I recall a buddy of mine playing Sacred 3, a subpar game by all accounts, especially compared to Sacred 2. During one mission, in which you’re expected to charge a rune while enemies attack you, he spent hours on it just because the enemies were endless. You were supposed to kill enough to charge the rune, use it, and move on. He sat there fighting these enemies for hours, gaining nothing but small amounts of gold, just to see if there was an actual end to them. Apparently not, but he had to go to work so he had to stop.
Procedural generation isn't an inherently bad thing though.
Maybe you should've not put any of that in there and put more time into the rest of the game...!
@@user-os7ec4dm8x it's like you're completely missing the point of procedural generation. they got to put more time into handcrafting the parts of the world they wanted the player to explore, without using an invisible wall.
I absolutely agree. I honestly get nervous when a game I am excited for has a main selling point of "this is BIG game." Right when they start doing this im like "O God here we go again." I think that the problem with procedural generation is that the tech is simply not there yet. People are not dumb (for the most part), we can pick up on patterns pretty quickly. The problem is that the minute you do pick up on it is the minute boredom starts to creep in. This is exacerbated 10x when the game relies on this technology way too much.
It isn't that the tech is there yet, because it's there - it's been there since at least Daggerfall, most of its landscape is procedurally generated. The problem is that it isn't actually very good, and the few people who actually sell it are more interested in either the math behind the coding or the illusion of endless possibility.
By contrast, devs and publishers like it because it means less time spent on coding or asset creation, which to them means more money saved that they can budget elsewhere (or magically put back in their own pockets). It really is a lose/lose situation for most gamers.
The core issue here is that while it's easy enough to fill a game with tons of stuff, it requires actual design to make it good.
It's possible to procedural generation right, provided you actively involve competent designers in the process. The simple solution that a lot of roguelikes have gone for is to build maps out of preset tiles, with the tiles being designed elements. The obvious issue with that solution is that players will very easily spot repeating tiles, so you still aren't getting a lot of mileage out of the elements that were made. Going a bit further than this means building design logic into the algorithm which can result in a less repetitive environment, but the player will still likely never see anything that wasn't specifically accounted for in the design in any case.
No matter what way you go about it, you will not wow players with hundreds of hours worth of remixed elements. The only way to keep players entertained for that long is for the gameplay to actually be good.
More often than not the only real value the player gets out of procedural generation is not knowing what's ahead. It adds replay value if the game was worth playing the first time but even then, once the player learns the patterns it's no different than repeating any other game.
They're just not using it right.
Hand crafted continents and landscapes followed by a few procedural generated dungeons for when the hand crafted dungeons have been plunged.
I mean, it works for Diablo 2.
The simple thing is, Bethesda does not know how to make the best use of Proceedual Generation tech and Starfield shows that very nicely.
Another way of using Proceedual Generation Tech and in a more correct manner, is the way that CIG is using it for Star Citizen, they create a planet with proceedual gen tech within certain perameters and after it has been created, they let their artists work on said planet to bring life to the planet, think of it as, painting on a newly created Canvas.
Starfields is.. Copy-pasta proceedual generated "stuff".. The way the two companies are using the tech, is miiiiles apart and Bethesda has no idea to to properly utilize it!
@@kerianhalcyon2769 it definitely hasn't been there since daggerfall lmao. that game was utter garbage, and its procgen was some of the worst that's ever existed.
Yakuza impressed me with its map. It's absolutely tiny but they make it feel different in each game. Seeing the 80's version of it, then early 2000's and then evolving throughout 2010's. Seeing the same noodle shop stay in place and not changing, you feel nostalgia for a place that never existed.
Are there any other games that do that? Same map but evolving with times and your actions from earlier games?
The most amazing example of procedural generation is Dwarf Fortress. The stories that its capable of generating is absolutely amazing.
Humans are really good at pattern recognition, so people start really noticing the repetition that comes out of procedural generation and everything starts feeling the same. As a developer you need to put in an exceedingly large amount of work into making it so that this doesn't happen in a players' playthough, but honestly that's probably more work than actually curating the entire experience by hand.
Exactly that. Wanted to throw Rimworld into the ring, but in the end Rimworld probably wouldn't exist without Dwarf Fortress.
CK series in a way but its much more "curated" but the options that can happen in a game sitting can be seen as "limitless"
The thing with Dwarf Fortress and Rimworld is that the procedural generation in a whole different scope from say... No Mans Sky or Starfields procedural generation. The colony sim style games work their magic by using procedural generation not just on maps but 'events'.
Starfield's procedural generation is 'here's a randomized map with randomized stuff on the map.'
Rimworld's procedural generation is 'here's a randomized map with randomized stuff on it, and here's a randomized weather event, and heres a randomized raid of randomized attackers, and here's a randomized illness on one of your pawns. Etc etc etc.' And on top of all that, the 'randomization' is actually curated by a storyteller in Rimworlds case.
That cascading effect of overlapping all these differing procedural generated things is what earns these games the name 'story generators.' And its how they avoid the 'mile wide, inch deep' effect we see in games like Starfield.
Never heard of this game until you mentioned it. Holy shit I’m buying it now! Thanks
@@FTORanger Bro, you're in for a trip with Dwarf Fortress. It's pretty overwhelming for a lot of people at first, and is one of those games you rely on a wiki for.
If you like the concept of Dwarf and can't click with it fully after playing some, I'd 100% recommend trying Rimworld out after that if you're in love with colony sim games. Rimworld is a great entry to the beast that is Dwarf. Same gameplay concept, different themes, narrowed down a bit, and has in game tutorials and tips lol.
Just, if you don't know how to play these types of games, losing is part of it. You WILL lose, so don't feel bad if you do! It's all part of learning how to play lol.
Also a very notable thing about roguelike procederal generation: very often they have set building blocks which are remixed and connected in different ways, but do have effort put into their design.
@@SawGudman I wouldn't even really call Path of Exile PG, there are MS paints-like maps and it always picks one of those and just makes the terrain on them slightly random. It is PG, but it's the smallest amount of PG you can use before you're just hand crafting maps.
@@SawGudman Starfield has like 12 enemy base layouts "garden" "mine" "large ship dock underground" and 1 forest and all of them have 3 layouts each :D
@@SawGudman "Daggerfall was huge and great, but got old very fast when the player realized it was just the same stuff over and over."(C) Todd Howard.
procedual generation is incredible, when it has restrains, guides and coherence. players dont need infinite content, players need the surprise and a meaning on these suprises.
Honestly the best take in this comment section. There's a time and a place for it, and when its done right, it can be a very useful way to enrich a game.
Fans just expected Bethesda to EXPAND on things they used to do, not dumb them down even further.
People wanted procedurally generated planets and dungeons akin to Daggerfall. Yet Starfield just plops the exact same prefabs on barren landscapes.
People wanted NPCs to have daily schedules like in Oblivion. What does Bethesda do? Removes daily schedules from NPCs entirely.
People wanted faction choices to matter like in FO4/Morrowind, but yet again we get Skyrim's "you can do everything in one go".
It's mind-blowing.
I loved the concept procedurally generated Levels but after seeing them for years only a few of them stand out and work. Mainly in the rouge lite genre.
The only reason this will work is when Creation Engine comes out and modders spend two years filling the content out so you don't encounter the same 'secret' mining facility or lab for the 15th time.
Hades uses it to great effect.
Yeah for roguelikes it works because repetition is the intended point. In games like starfield, there's no reason why repetition and repetitive environments should ever be viewed as a good thing.
@@headkicked What did surteabag8652 just say?
That is because procgen cannot stand out by definition. It can only remix assets that only exist, and in most games that means you don't really register the actual layouts of said assets, you notice the assets themselves.
Procedural generation isn't a bad idea, even in the "bad games" Josh mentioned. It's just a bad execution. The whole concept is basically "done dirty" by how it is used: a simplified randomization of pre-created assets. A few different tables at best like "biome", "in-door assets", "out-door assets", "loot table", "enemy table"... mb something like "miscellaneous". So it obviously generates dead-on-arrival, sterile environments.
And it isn't that hard to fix. Add more conditions, add more criteria for picking assets, add memorization so that areas won't be a copy of each other with 1 element changed, add statistics analysis so that your auto-generated level would be on par with manually crafted ones. I truly believe that it isn't that hard, really. It would just take time. More time. Probably more people involved. And more testing to polish it. And all of those things mean "MONEY": time is money, adding more people to a project means money, testing/polishing something requires money...
And a company as a whole doesn't really need a "good product", they need a product they can "sell". So instead of a tool (talking about procedural generation as a tool here) that could makes game GOOD, they just use it as a tool that allows them to SAVE TIME (=SAVE MONEY). That's really all there is to it.
Something like the orc captains in Shadow of Mordor.
Dwarf Fortress is the worst possible example if you're trying to argue that it's "not that hard". That game has been in development for almost 20 years, and it is massively complex, with dozens of incredibly intricate systems interacting in ways that the devs have to make sure doesn't break the game.
That is a very hard game to make, and would only get harder with a large team because of the coordination required between disparate systems. In general, good proc gen is so much harder than "just add more conditions lul".
In terms of procedural generation approaching some kind of ideal that AAAs could never touch, I always think about Dwarf Fortress. Its whole world and all the characters and the entire written and unwritten history of everything and everyone are generated. In the game's fortress mode, this mostly affects you with the megabeasts that occasionally invade your fortress and the visitors that loiter in your taverns, the personal histories and skills/knowledge of the migrants who move in and become your citizens. But also with things like your civilization's technical knowledge (how much of the invisible "tech tree" has been opened up already, with scholars and a library needed to get more), news from trade caravans about recent events in the world, dances and songs and the instruments needed for them created over the years by bards and musical troupes, engravings on walls pulling directly from historical events.
Then in the adventure mode (which was temporarily cut from the game's steam release but will come back some day, maybe next year), it was a lot more jank, but as your adventurer party roams the world, you could ask any random character out in the world about any character, location or rare artifact your party has ever heard of, and usually they probably won't know what you're talking about, but other times they could point you in the right direction and help you find your way with the (again, janky) quests. It always fell short with how limited the dialogue from the npcs was and is, though, unfortunately, as that's been one of the game's weaker spots for a while. You can ask a dizzying number of pre-formatted questions with the subject swapped out, but what you get in return often just ended up being "it is inevitable" or one of the other go-to phrases.
There's definitely so many more years and updates needed before most of DF's features feel better, more cohesive, more playable, since their ultimate design goals are ambition incarnate, but the devs still put a hell of a lot of work into the game's systems as it is, and plan to put in a hell of a lot more, and it's magnitudes more time and effort (and money) than what any AAA devs or their shareholders would ever be willing to invest in one title. But it gives me hope for better and more nuanced implementations of procedural generation in games, just by the merit of it existing, even if it is a one of a kind game.
When you think about it, the entire real universe is procedurally generated - just with incredibly complex conditions and rulesets: given identical conditions, thr ro k of a cave system will erode the same way every time. But of course they are not identical conditions: the rock has different makeup; the water varies in quantity, and ph; the caves all started at differing times... Planets and suns form by the same rules - they couldn't be more different.
@kekker_3362 but: A) it was still made; B) people loved it. Of course making such a system isn't as easy as to describe it in a few sentences in UA-cam comment section. I was not implying that)) But it's also not a ground-breaking invention, that requires dozens of geniuses to achieve.
And also saying that it's harder because of number of people involved is really funny, considering how big game-developing teams are nowadays. Have you ever watched game's end-screens? Hundreds of names! Adding a few more - and in most cases you don't even need to add, because they would probably be the same people already working on the game! - won't be anything new. Coordination is a task - I can agree on that, - but it is already tackled. You can't use it as an excuse for not doing something. It's kind of more ridiculous than spending more money on a project.
I like how Subnautica mixes procedural generation and design. The map and most landmarks are fixed points, and some creatures are placed with pathing and AI. However, the resources and most of the creatures are procedurally generated based on the biome. It makes each playthrough different and keeps the world feeling alive, and it doesn't mess with the progression or shatter immersion.
The one thing that really throws me for a loop in Subnautica is how the life pod spawns at different points in the safe shallows on each playthrough. I'm always lost at the very start
Oh, I didn't even know that, and I love that game! The thing is, the game is so huge, yet its all made by actual people. So I didn't even notice there was a any procedural generation going on at all.
@@tenyokensekia8088
Because Procedural generation isn't inherently bad.
Some of the best Dungeon crawlers are procedurally generated.
You can make it work, it's just that Bethesda decided to use procedurally generation to replace everything.
I'd love a modern RPG that did what Diablo 2 did.
The procedural generation is not the problem, it's just the people, and it always has been just the people.
@@citizenvulpes4562 I never said it was inherently bad, and I totally agree! I love it for dungeon crawlers too.
Yeah but the progression is still the same. The only difference would be the time it takes to progress.
also the beauty of nothing comes when you don’t really expect it or you know how you got there. my problem with starfield is it’s like knowingly taking a loading screen to just sit in the middle of nowhere. it takes away the magic and doesn’t feel like you stumbled across this vast landscape you just loaded into one
Since Starfield I decided I'd skip it and just play Morrowind. Vastly smaller, but my previous playthroughs I avoided whole areas of the map and whole quest lines. Its interesting and I have saved a game out in the wilderness and lacked tools to get an NPC escort home. Also Kerbal Space Program is a game about dead empty planets with nothing there. Its amazing once you get into it and start building ships that can reach new worlds. They made the mechanics of travel the key, and its an exploration game. Finding anything weird at all is something you run to take screenshots of and show a friend.
I think part of the reason studios go for bigger quantity goals is to have “objective” metrics to brag about. They love to say the numbers of planets, the area of the map, the line count of the dialogue (dialogue is something starfield specifically bragged about in ads)
This would've been VERY impressive in 2010, but in 2023 we've already been fooled enough times by massive worlds with nothing to do in them. I honestly can't believe there were people excited to hear about the 1000 planets.
It's also a fantastic fit if you wanna sell a subscription service like Gamepass and cannot rely on the good old 70 bucks per game sold on release kind of business model anymore. Probably why the new Forza game has you level up each of the 500+ cars individually.
I hope the whole subscription crap isn't gonna result in the next race to the bottom akin to what happened with the mobile games market.
I think they're just behind the times a bit. There was a time when open-world + sandbox + big was a massive novelty, a technical achievement if you could pull it off, and something players thought they wanted (as opposed to being 'on-rails', rooms+corridors etc). And people did. Then they saw that maximalist interpretations of that aren't fun and aren't as good as they thought they would be, so they've settled on wanting something smaller, with a good, hand-made, tight central plot, some curated side quests and plot lines and a sandbox for those who love the world and want to go off and live in a tree surrounded by modded-in s-xbots. The execs just haven't caught up.
Atomic Heart is a good case study. It was going to be a big open world sandbox originally (with a plot), but pivoted to a much tighter scope and benefited from it. There were a few grumbles in reviews about the wider world not being as big or populated as they expected based on some early discussions, but no one really cared enough about that to take more than a point off the rating. If you can't make an engaging big world, don't waste the money and time because players will not appreciate big for the sake of big in 2023.
@@CidGuerreiro1234
Exactly, the game is outdated, despite coming out this year.
It makes sense when you remember that Starfield is built in the same engine Morrowind was built in.
Even with a 1000+ "planets".. Starfield is still smaller than Star Citizen..
Which of these two games has more quests?
Game A:
1) Speak to the townspeople to gather information and make a plan.
2) Gather items you need for your fancy plan from three different enemy types.
3) Carry out your plan to raid a big enemy base and defeat the big bad within.
Game B:
1) Deliver 6 wolf pelts to a leatherworker.
2) Deliver 8 fire shards to a blacksmith.
3) Deliver 12 spider threads to a weaver.
4) Deliver 5 spearheads to the general.
5) Deliver a baker's dozen eggs to a chef.
The kneejerk answer is that they go 3v5, but a careful inspection shows that Game A has three quests, and Game B only has one. Even if it had a thousand such quests, it would STILL only have one quest. That's what these game devs aren't seeing.
And all that's not even considering quality, just quantity. A game with a thousand quests loses to a game with 3, on QUANTITY of quests. Quality is judged separately, but history has shown that the 3-quest game usually wins on that front too.
Josh said he'd prefer a well-made 10-hour game over the 1000-hour procedurally generated one. I'd go even farther and say I'd prefer a well-made 10 MINUTE game. Which I can back up with a history of playing short little flash games on old sites like Armor Games, Kongregate, and the like.
Nah mate, my knee jerk is that it's 3v0. Deliveries aren't quests, they're side jobs. Ways to grind up a little extra cash, or your first minor task before your ready for the big leagues. If fetching some materials is a quest, then the word has lost all meaning.
@@lawaern3474 That's another way to look at it. Personally I don't mind being generous and calling it 1, especially since that leads to "at my most generous that's still atrocious". In moderation and spread out between more awesome moments, the rare kill-x-things quests don't bother me at all. But I also completely get it if even that much bothers other people.
@@riluna3695 They don't bother me too much either, heck one my happiest gaming experiences (gothic ii) had a kill-x quest at the heart of it. I just dislike the gaming term "quest" being used for anything an npc asks you to do. When I went out to get some bread from woollies this morning, it wasn't a quest. It was an errand. But when I crossed Gielnor to gather rare and absurd ingredients for One Small Favor, it was a QUEST.
But, I firmly believe errands and other minor tasks have a place in adventure rpgs. But we play these games for the quests, not the errands. Calling them quests just gives game devs excuses to pad their advertised numbers.
@@lawaern3474 Ah, I see what you mean. Yeah, I agree completely. Best to call it what it really is.
That also reminds me of an old favorite cartoon :D
"So it's not really a mission, it's more like an errand."
"Well missions are just errands that....involve death."
"So like...a funeral...would be a mission."
"...You think funerals are an errand?"
You speak wisdom sir, please continue cooking ;D
I'm a senior game designer who's been working in the industry for over 8 years, and I can confirm your suspicion that those who advocate for procedural generation are more often than not directors or executives with little understanding of what provides genuine value in a video game as a product.
This thing about limitations is something i always talked about. So many awesome things in old games were the result of technological limitations of the time and the need to find a creative way around them. Today they have the technology to run wild with any crazy idea but they can't come up with ideas because there is nothing forcing them to think outside the box. The easier the work is, the more formulaic the product
Nah, still happens in indie games. Public traded companies just have absolute inhumane monsters as execs and suck all fun and innovation out of video games. Just imagine what a studio's burn rate is with 400 plus employees. They have a hard time justifying taking risks. So they just copy what worked in the past.
I continue to be impressed by the code that was OG Pokemon, bugs and all. So much was crammed into something smaller than most .pngs.
@@1IGG And the other part that a lot of folks don't like talking about is that consumers don't like risk either. Even if a change to a franchise is good for it's health in the long run.
@@1IGG But indie games do have limits.
Limits in their budget go super far.
They often can't have a huge team of artists to make something on par with AAA so they are limited to something more stylized. The budget for sound effects will limit their ability to compete with Call of Duty and Doom 2016 for satisfying sound effects and gun feedback. And engine + creative limitations are still a thing, you don't have teams of hundreds of people to figure out *how* to do something you may want to, so you gotta figure out a workaround.
The only limitations it feels like retro devs had that moder. indies don't is hardware limitations. Otherwise, Super Metroid and Castlevania SotN weren't made by any more than 12-30 people depending on how you count musicians.
I recently left university with my masters and started at an executive level in a company (not gaming) and I see what you mentioned in both my colleagues and higher ups as well as finding it deeply frustrating
It’s all about making numbers go up at the expense of a good experience that actually makes sense and engages the customer
oh and in your last sentence you accidentaly described 90% of WoW players who raid. (me sadly included)
then start your own business and see how far you can go without covering costs, its unfortunate but that is just how it works
@@yonosepe7491 This is a meaningless comment. Lots of companies make *good* products while covering costs. It's greed pushing these companies to forgo average customer experience to chase ever-increasing profits. How are you having trouble "covering costs" when you report record profit to your shareholders? Those are diametrically opposed concepts. Bethesda in this case dumped hundreds of millions of dollars into Starfield. It wasn't a lack of funding or cost covering issues. It was poorly using hundreds of millions of dollar and *years* of workhours mismanaged because they bought their own hype and several senior members working on these games have a mentality that because Skyrim and its 12 versions or whatever sold amazingly that their ideas are above reproach and criticism. They were so wrapped up in wanting it to be the absolute smash hit like Skyrim that they just made Skyrim again but worse, and now Starfield's cultural legacy will be highlighting how truly out of touch the leadership at Bethesda Game Studios has become. They don't want to *work* for the accolades and sales figures, partially because of Todd's trauma with the Morrowind crunch as they teetered on bankruptcy, and partially because many of the senior members think they deserve it, despite their formula being *decades* old at this point and pretty tired on its own. There's a reason more people have played modded Skyrim than people who have finished the main quest lol. At this point they're better at just making models and worlds that feel good to wander around in.
Nobody is talking about an inability to afford the work needed to produce something good. That's not even a factor in this equation. The number couldn't be going up if the costs were outweighing the profits.
@@yonosepe7491 There's definitely a middle-ground between Not Keeping the Lights On, and Make Net Profit Go Up I Don't Care How
A designer knows he has achieved perfection not when there is nothing left to add, but when there of nothing left to take away. -de Saint exupery
Doesnt sound right. Its extremly easy to fall under the impresion that there is nothing left to take away.
"I didn't have time to write a short letter, so here's a long one"
Starfield somehow manages to be a game about exploring space that somehow manages to lose the wanderlust. Once you have played 3 hours you’ve played it all.
it's like they learned nothing at all from no man's sky
Agreed.
No man's sky is the only good procedurally generated space exploration game.
Starfield was outdated by a decade on release. But people will still meat-ride it due to sunk-cost fallacy and wool-covered eyes.
@@jayson9999ful no man's sky is hot garbage
Its a Space Exploration game about two guys with guns throwing space magic and bullets at each other.
That is so not true at all 😄
I put 100 hours into Starfield purely due to the sunk cost fallacy. I tried to convince myself that the bits of fun i was having were enough when they clearly weren't. Meanwhile, i beat all three endings of Armored Core 6 in the same time and still think about how much fun i had playing it
Armored Core 6 is one of those games that just invite the 100%
Imagine if starfield has space combat and ship building as fun as armored core 6 but bethesda put lowest effort to their game mechanics
446h on bg3 and currently i'm still on 2nd playthrough.
this is truly mindblowing. there are so many to discover even on my second playthrough, and I'm on the "good" side on both. the game is quite buggy in Act 3 and Gale is accusing me of making an agreement with a person i had yet to meet, but the game as a whole is still incredibly well made. it is a gigantic flawless gem with a few scratches that they are still polishing.
I had the same issue in BG3. Also for some reason I have an option to ask a companion what happened to our romance even though the romance is full on ongoing. Minor bugs and AI movement aside, the game is really deep and i love it. Just like a great RPG is meant to be.
I also am seeing how buggy the game is in act 3. Hour 150 in my first play through, and I've had the same Gale issue
Yea a couple of times my companions seem to know more than I do but that happens in real life too so I let that bug pass. Awesome game. It’s turning into my fav western turn based game of all time
Tooo bad it's a Tumblr universe
@@arynasabalenka3173 what the hell is a Tumblr universe
The GDC presentation for Into the Breach was eye-opening to me: that game has no procedural generated levels, despite being essentially a rogue-like puzzle game which, on the surface, would be perfect. Instead, they realized it would be faster to make hundreds of challenging and fair levels themselves, rather than perfect an algorithm that might not work (or probably not be fun). That same reasoning could be beneficial to a lot of games.
This.
To make proc gen to be good you need to put more work than actually making the content.
Hades is that way too, the maps are set, and the encounter variety too, so it shuffles between set designed rooms and enemies/upgrades.
Into the Breach is such an amazing game. A SHAME it's not more known around.
@@santiagorios2083 I didn't realize that for Hades, but now that you mention it: the Binding of Isaac is set too. It proc gens the map layout based on simple rules, but the rooms themselves are hand-built (with rewards scrambled by seed). FTL and Slay the Spire also only use limited proc gen.
@@santiagorios2083Hades is a game that feels like devs play games not as work but as passion and hobby. All rooms, despite being able to be crazy hard, are fair. Even the shittier ones.
The problem with how a lot of AAA devs use Procedural Generation, and Bethesda are kings of this, is that they use it for the sake of it. It’s a marketing bullet point, rather than part of the core game design equation. You look at something like Dead Cells where part of the challenge is not being able to memorise the level layouts because it changes every time you die, and you get a design that was built around Procedural Generation rather than one that has it bolted on to the side.
To make a comparison to a more similar game: No Man’s Sky. It’s a game where every environment is made with PG, but the entire point of the game is exploring these endless new planets, and each one is a complete planet with its own biome and flora and fauna. It was a core part of the design, and the game is better for it because practically nobody has seen the same planets as you. NMS’ PG is used to give you an endless amount of exploration and discovery.
Then you have Starfield where it’s basically used as a clunkier and less interesting version of Skyrim’s caves because it was just an easier way to add a large amount of content to the game rather than making content that’s worth experiencing for the experience itself.
If Procedural Generation isn’t a core part of the game design, then what you generate with it has to be something incredibly entertaining or it’s going to feel like an all-you-can-eat buffet of flavourless slop.
Nothing is better than that feeling in Elden Ring when you enter a new area and know that something important / cool awaits you.
I don’t think any other game filled me with as much awe and wonder as consistently as Elden Ring did except for maybe Outer Wilds (but that is a much shorter game).
A reused asset.
another important/cool tree spirit that you have seen 5 times already
@@fartloudYT I will maintain that if Elden Ring hadn't been a Fromsoft game, it never would have gotten the reception it did.
@@mechanomics2649 Is that why it made Fromsoft much more popular than they were ? Elden Ring got beyond mainstream and let me tell you the amount of people who played it who hadn't heard of fromsoft is absolutely enormous. It is an insanely high-quality game and them re-using a boss 3 times won't change that.
I think procedural generation can work occasionally. Hades and Remnant: From the Ashes both had a bunch of curated rooms that would get laid out for you. In Hades, it’s before each run. In Remnant, when you started the game it was procedurally generated so you’d have a unique map to someone else, and that was a cool idea, but every area was made and then connected to each other. You’d have a boss fight in one play through that wouldn’t show up in another and that was cool. But in Starfield it’s just a bunch of empty planets for what?
Looks like someone hasn't played Starfield.
Look tbf its a bethsoft game, its fallout and skyrim elements put into a space genre game. But still your comment proves you haven't played the game. So cope.
Dark tide has a similar thing as well. Premade rooms all procedurally generated together. It's great
I dunno about Remnant, but Hades is not procedural generation. There is a preset selection of assets for each level, a preset layout and preset potential enemy spawns. It randomly selects from the assets available, it doesnt just CREATE new layouts from the assets.
@@ac3d657Cool, you totally dunked on a pointless comment. Now dunk on the points from the video. Or, you know...cope.
@@ac3d657star field blows and I have played it. So cope more. Game is total trash.
a mile wide but only an inch deep
the thing is its not even that wide
Sometimes the 'nothing' can be used for great effect, it can be Poignant and a breather for the player. The long ladder from snake eater, the quiet paths in dark souks that connect one area to another, the open areas in shadow of the colossus to emphasise how big the boss's are. But big open nothing for the sake of nothing is just pointless.
I think it would be beneficial to distinguish between a randomizer (Hades) and procedural generation (Starfield landscapes). When you take a big pile of handcrafted elements and draw a handful out of the bag and let the computer string them together randomly thats for of a randomizer than procedural generation... procedural generation creates topography, forests, puts in animals and sometimes enemies.
Still, making the distinction does not tell whether the game is good or not. Minecraft and Terraria's worlds are fully procedurally generated but they are great games and they wouldn't even really work without the procedural generation. But that does not mean it works in games like Starfield.
Yes, and if it was implemented on Skyrim it was definitely checked and edited by hand, which can take longer, so possibly not even used. The Starfield moonscapes are obviously generated - planets and moons obviously can't be made by hand. Starfield has large hand-crafted bases, but the 'open world' beyond is really just tacked on, as it's expected, but it's almost irrelevant, hence making moon buggies was not worth it for them, or too difficult.
They couldn't have Skyrim's great explorable open world and space travel to hundreds of planets. Considering how dull the space travel is, they really should've had one huge planet, with maybe a moon, and an agile spaceship to fly and land to explore (a bit like #HaloInfinite, but far bigger, and better spacecraft than the Banshee) - but that's all technically beyond Bethesda, who'd rather use the processor power on the 1000 looted items you have in your pockets or have scattered everywhere.
The wonderful game Enter the Gungeon also did that "some units out of the bag" design to great effect!
@@miriamkapeller6754 The only way to create planets is to use Procedural generation
@@goatlps Having only one planet would have drastically limited the location diversity, biome diversity and the worldbuilding.
Also it doesn't have the space exploration vibe, you won't be discovering new systems, you won't be finding beautiful planets to build outposts...
The space travel is dull ? Even in No Man's Sky you need to "fast travel" between solar system.
I was so impressed with all the marketing around Skyrim's "Radiant Quests". I can remember being in Skype and Teamspeak with all the mates I played with online, we were all so hyped about it. Infinite quests, we could just keep playing FOREVER!!!, no need to worry about running out of things to do.
Then we actually got our hands on the game, and while still a revolutionary game for it's time, all talk of "Radiant Quests" being the next big thing died immediately.
It almost seems that every couple of years it gets brought back up again, that technology has progressed and "this time it will be different". But nothing I have seen has changed my opinion on em.
It's just an another ridiculous example of the way Bethesda are stuck in the past and coasting off past success to the absolute extreme. They're repeating the same misguided adoration for procedural generation long after everybody has seen it for what it is.
thing is Radiant Quests arent "new" in Elder Scrolls actual Radiant Quests started in Daggerfall.... and... Daggerfall being Daggerfall.... YEAH...
some people have even gone about speaking that Starfield uses Daggerfall as its precedural generation process which may or may not be far from the truth as there are only 10-13 dungeons in that game that are hand crafted the rest are all procedural generation with expected jank
There's a few marketing buzz words that just seem to enthral gamers for no logical reason. A big one I can also think of is a game having a "10-year plan" All you have to do is look at GTA v to realize that maybe a 10-year plan isn't actually the best thing
Yeah, it's bizarre, because Morrowind was such a hit precisely because it got rid of Daggerfall's procedural generation, and everything was done by hand!
I told a buddy I did everything in skyrim and he goes "there's literally infinite quests" and like, technically yes, but it's really just the same thing over and over again.
5:50 YES, absolutely! I used to become incredibly hyped whenever a new "massive sprawling open-world RPG" was introduced but I learned my lesson.
Nowadays, I try to temper my expectations and adapt the "wait and see" approach. I'm still interested if I see these types of games being announced, and I'll follow up on first news and reviews, but I wait until the game is released and then some more to see the first user reviews before I make up my mind whether I'll buy/play it or not.
Daggerfall is a really interesting game in procedural generation. The game does repeat random quests insanely and the dungeons were randomly seeded when the game was made, minus main quest dungeons, but still the game has this really addicting gameplay loop that makes it satisfying to play.
I tried playing the Unreal version of Daggerfall a few times.
I really would love if they could make a game with modern graphics and controls but keep everything gameplay-wise that made Daggerfall and Morrowind great.
I wish Bungie had done that with Marathon as well. The Marathon universe with Halo Infinite graphics and gameplay would be a riot.
@@atraxisdarkstar ... That's Halo. The only thing from the Marathon series really missing in Halo is Durandal.
Indeed, the gameplay makes Daggerfall work. Imagine this: port Skyrim or Starfield BACK to Daggerfall graphics. It will be clear how much gameplay was lost.
The point of procedural generation imo is to give a sense of scale quite easily without needing to produce alot.
Daggerfall does this pretty good because the worlds as a whole are massive but the important elements are all hand crafted.
But then again small worlds but with more detail is better than bigger worlds. For example the yakuza games have small open world but very detailed and each street corner is filled with life (kamurocho especially so is tiny but high quality).
All of the content in Starfield is handcrafted, only the landscape is proc gen
@@ni9274 Repeatedly telling a lie doesn't make it more true... just saying...
Starfield is a game with 100's of NPC's and no characters.
I strongly disagree. There are infinite soulless NPCs in Starfield but there are also a lot of interesting and well written characters.
@@thebaffman4898 A few diamonds in a sea of poop
Interesting and well-written are not what's on offer in Starfield@@thebaffman4898
@@OnlyGettyYou can say that of many many games. I don't get why people are so mad at Starfield. I mean, I understand the disappointment, I am disappointed myself, but I'm just trying to enjoy the good bits instead of shitting on the game at every chance I get. And there's A LOT of good stuff in Starfield. I am 50 hours in and I landed on maybe two planets that were not related to a quest. I barely scratched the main campaign because of the amount faction and side quests I got involved in. I just avoid the boring stuff. Nobody is forcing you to explore every single cave and talk to every single NPC that says nothing but useless boring dialogues.
@@thebaffman4898 Let me help you with that. Compare the game to Skyrim. How much has changed between the two games? The answer is: Surprisingly little. Starfield isn't a space game. It's a Bethesda game that takes place in space.
Or maybe it would be more appropriate to say, Starfield isn't a game. It's a game engine with a story attached to it. Like Skyrim, life will be brought to the game through mods.
I said in our whatsapp in June that starfield would be a mile wide, but an inch deep. A couple of weeks after release, we had a news article with that exact headline. Every time a dev keeps pushing the massive scope of a game world it ends up the same way
I mean, you're not exactly a prophet for saying so. Plenty of people were saying the same thing and it's not hard to see why.
1:22 I think Josh is wrong here. By now gamers might have somewhat caught on, but I still remember how massive the hype for No Mans Sky was, how often I was told that it’s basically infinite and thus going to be the best game ever made.
Or even the “hour per dollar” calculation, as in how many hours you can play a game and comparing that 1:1 to money spent. A massive amount of people believed this to be a good way to judge if a game purchase is worth it, many still do.
So it’s not just executives who want these massive games with slop content, a lot of gamers do too.
A great use of procedural generation is in Dwarf Fortress.
The reason it works is how deep it ll goes, as the end goal of the devs is to simulate all of reality.
For example, when a necromancer attacks you, it will not be some random enemy, it will be a person with a history, what they life was like and what lead them to learning the secrets of life and death, their rise to power and eventually the moment they attacked your fortress.
If accompanied by, for example, and undead dragon, the beast will also have history of what they did before being struck down and eventually being revived by the necromancer years later.
This depth is also applied to civilizations as a whole, scholars, heroes, artifacts and many other things.
In a different game it would likely be a randomly generated necromancer, with a randomly generated dragon by their side, with no further depth.
Procedural generation works best in games that are designed around it, like Deep Rock Galactic and Dwarf Fortress (funny that both games are about dwarves).
Bethesda, on the other hand, used procedural generation as a tool to additionally inflate their world into a bloat of blandness.
@@sujimayne Yep, procedural generation has it's uses and now that executives in these big companies finally found out it exists, they use it for things it is not designed for in order to inflate their games and save on time and work it would normally take.
It's used to generate worlds/levels in certain games for a reason. Those games use it more as a backdrop rather than a deep world to interact with, making it a perfect solution. If it's used for characters, lore, etc, it needs to have a lot of depth, which requires much more effort to accomplish, than doing it the regular way.
The moment I realized procedural generated quests where a handful of nothing was in Oblivion after I became the Listener for the Dark Brotherhood and started getting those weekly meaningless contracts.
Speaking of limitations. I have two examples. One recent and one a bit older.
Dead Island 2 does not use traditional shadows. It is a trick. It saves a ton of resources by making the shadows feel and look like they are true shadows casted by objects but they are not, they are literally drawn. It's so good that only in specific places and very specific camera angles can you actually tell.
The other one is Xenoblade Chronicles 3. Arguably the most beautiful game on switch. And there it is. Switch. It manages to look ridiculously good while being limited by hardware. It doesn't render all the flora around you at all times and saves resources that way but still gives you the impression that everything is being animated all the time. It holds stable 30 fps while 7 of your characters are doing ridiculous over the top stunts fighting 5+ enemies doing equally ridiculous stunts full of particle effects.
We need limits to find our true abilities.
I completely agree here. One of my guilty pleasures, I suppose, are the Assassin's Creed games, but the plot is just a small percentage of the entire game. I played Odyssey for like 150 hours and spent like three quarters of the total time clearing out the map that is littered with the same two or three basic activities. I'm so done with games claiming to be hundreds upon hundreds of hours in playtime, but once you play them you quickly realise they only achieve that time by having a massive, mostly empty world with about three or four activities, which are being repeated time and time again.
Deep rock galactic does procedural generation really well. They put the generation on a track with some set rooms with random variants in actual geography and random gameplay elements. You eventually notice repeating room designs but it takes a WHILE. Like easily 50-100 hours before you notice constantly if you've been to this room before.
Rain world also uses this tool well. They randomly generate some creature animations with some animations being set. It works well because the creatures don't resemble real animals so our frame of reference for how they should move is limited. It nails that feeling of a living world by far the best.
Agreed! Excited for their new games too :)
Yeah, procedural generation is just a technique, it's not good or bad, it depends on how it's used. There are a lot of games that uses procedural generation very well, you can't just say "procedural generation sucks because in Starfield in generates a lot of nothing".
Yeah, the real question is what the procedural generation is used for, and have good is the algorithm. Minecraft is a good example of a great algorithm, to this day I enjoy just strolling through and seeing what landscapes generate.
@@85blutch Name one.
@@mrosskne Dwarf Fortress, Minecraft, Terraria, The Binding of Isaac, XCOM 2, Dead Cells, Hades, Rimworld, Deep Rock Galactic, these are all great games that use procedural generation in some way.
I remember josh going over this exact thing in his Castlevania review, where the limitations of the engine are what pushed them to get the most out of the hardware.
i definitely felt this way with elden ring's dungeons. The rest of the game was fantastic, but the dungeons felt very bland unless they had some sort of puzzle to them (like the one where its two identical dungeons repeated on itself, that was cool), and the bosses were often ones I had fought a good few times before. I would have preferred much less dungeons like those ones, but more of the altus plateau secret passage one, since that felt really unique and had a lot to explore.
Hmmm visualy maybe but never have i though they were bland or uninteresting
One of few games i had to see everything cus i wanted to
I had the neeed to see it all
Elden Rings dungeons werent procedurally generated tho, its just that designers focused more on open world and smaller important sections where main bosses reside. Dungeons are just these secondary mini experiences inside grander scheme. Also I wouldnt say they ALL were that bland.
ER is a covided game, when from took their time AND FOR 2 YEARS already are trying to fix it.
I think that’s because you’re not meant to go to all of them in one playthrough, which is a shame.
Elden Ring was definitely that game which I was glad to finish and put away after nearly 120 hours.
It was right at the edge of overstaying its welcome because it was really pushing its design limits at that point and becoming repetitive.
I did come back for a Seamless Co-op (mod) + Unalloyed Modpack playthrough with a friend and it was a blast.
6:00 I'm impressed with the math behind it. Procedural generation is only successful when it involves secondary systems. For example, the procedurally generated enemies in Shadow of War work out just fine. Making quests involving them would feel natural due to how the system works.
I really like use of procedural generation in strategy games like Age of empires, where you select theme of the map, but the placement of resources is always a bit random, it prevents meta gaming. In comparision in starcraft you can just learn the map and use strategy best for that map every time. I like the feeling of little bit of unknown going into the battle and real fog of war. I now it can get unbalanced a bit, but thats just like in real life and you gotta adapt to it. Best use ive seen is in little strategy game called Tooth and tail.
Random ressources spawn in RTS games can be tricky because certain players may spawn with a whole lot of ressources in proximity whilst others will struggle to build their economy
@@KalashVodka175 You always spawn with a gold patch and a stone patch over a set distance in AoE, only the location of them changes, so you still have to find them first.
@@fus132 I love that from aoe4, even tho sometimes it is at my disadvantage, it really changes the feel from one game to another
I can tell you don't play Starcraft
@@KILOBify I used to play for years. That is why I made the comparision.
Everything you said about working around limitations really made me think of things i hadn't before, thank you for the inspiration!
There was a time in my life when 20 hrs game was finished in like two evenings. Now that time streches to almost a month. Yeah, times are changing
For you yeah, younger people with more time do and will always exist though and what they want isn't irrelevant either. I'm glad bg3 is a 150 hour game for example, it's an amazing game. I would've been pissed if it was
"Flavourless slop" - In a year of brave attempts, there has been no finer way of describing Snorefield. That's exactly what it is; slop. I sometimes wonder what must have been going through Todd's mind in the weeks leading up to the release of this haha game. He must have been absolutely shitting himself, he must have realised it was a giant -tasteless- turkey but after all that time and all that money, he could not back out. I almost feel sorry for him until I remember that they worked on this instead of the next TES. Criminal.
Im currently playing bg3. 80 hours in. Just reached act 3. Been playing with a friend who has finished the game at least twice and already started other playthroughs and even he said hes seeing new stuff based upon my choices.
Thay speaks volume to me, because ive loved this imperfect playthrough and cant wait to see how things turn out when i plau through it again.
Seriously, this game is a gem
you get new options even when using different characters, it's amazing
did you fuck the bear yet?
You hit the nail on the head. Procedural generation was touted as a selling point for Skyrim before its release. Radiant quest had me excited. But when the game came out and I experienced those quests, after the third one I understood they were basic fetch quests with no impact. I've been turned off from the idea ever since. Bethesda seems to keep leaning into them though so I suppose someone likes them.
All of the quest in Starfield are handcrafted
@@ni9274 maybe faction quests or the main quests, but everything surrounding those is proc.gen, ands the fetch quests are definitely proc.gen
Procedural generation is not what people buy their games for. They buy it to mod it.
@@iplyrunescape305 I'd agree with that if it wasn't for the massive amount of obsessed fans who buy it day one. If they bought it for mods, they'd wait until the modders had a chance to fix things and turn it into an actually worth buying game.
The sad truth is that there are a lot of people who buy Bethesda games because they remember their glory days of stuff like Oblivion and Fallout 3, and because despite its flaws Skyrim is still a good game with only the starting of "Modern Bethesda's" lack of effort. It's the same reason people keep buying newer Nintendo games despite their decreasing quality and still praise them in spite of the flaws, or why people keep buying Halo, CoD, or Assassin's Creed. Some people refuse to give up on developers or franchises because of nostalgia for what once was.
@@Gamer3427 you are right, it's just since Skyrim the only living legacy is that it's highly moddable. No one goes like "damn I love playing through the thieve's guild and become the guild master in like one hour" or remember anything about the story. Same thing in Fallout 4. 76 required more effort to "fix" from Bethesda since you couldn't mod it. Starfield just more of the same problems, people are getting weary of this.
The diehard fanboys yes, buy it reminiscing Oblivion and Fallout 3, but they drop the game pretty quickly and only come back for mods. They probably defend it still by saying "just get a mod" to "fix" problems
Procedural generation generally works best when the environment is the variable on a solid set of systems and it's not too extreme.
Eg: POE / Minecraft / Deep Rock Galactic / Factorio:
Maps are slightly varied, but the foundational mechanics are well conceived and consistent. Those games also have very limited focus on story and they're more about looting and group play.
Warframe does the same with its tile sets, and it works perfectly for that type of a game. Shadow Warrior 2 used a similar system as well afaik. While it was overall pretty solidly put together, the game did suffer in my opinion from having less distinct and unique level design.
It's not only boring, it creates a sense of emptiness, loneliness, inauthenticity and the list goes on. Procedural generation is decades off from being on par with meaningfully curated content.
Procedural Generation has its benefits, but there needs to be meaning there. Starfield has empty space that isn't backed by story or gameplay. If the whole game took place between a unique planet and unique moons that replaced other planets you'd have a tighter experience in theory, and the game would still be empty.
@@Pangora2 essentially procedural generation is a tool, not a finished product.
it only works in stuff like roguelikes because there the core gameplay experience that relies on doing multiple semi-short runs actually gets enhanced by having a random variance
The end of Holy Grail also works as both a cost saver, they avoid the production cost of a big battle by having the police arrest everyone beforehand but also ties into earlier in the film when the reporter is cut down, we see his widow with the police.
The primary reason us old timers love baldurs gate 3 is that we played bg1 and bg2 20 years ago and creamed our pants even knowing it will take us years to beat it
While I'm in your camp, I don't think the general public is. My wife grew up without PC gaming, or dnd. She's absolutely hooked on BG3.
@@carsonspears8568I’m 45 years old and never played a crpg before. I’m now 50+ hours in bg3 and I’m loving every minute of it. It’s opened up a crpg rabbit hole for me now, lol.
You hit the nail on the head around the 5:53 mark. I was excited for Starfield and played it a lot when it first came out(and still plan on finishing it, or the main story quests at least), but the more I played, the more I saw how randomly visiting planets and systems were basically pointless. Sure you could set up outposts and farm resources, or you could do what I did which was loot all the weapons and armor off enemies, vendor them wherever I could, and then just... buy resources from a vendor because it was just easier and you didn't have to go out of your way for it.
Completely takes the point of resource farming out of the equation. The weapon and armor system also feels like a half-baked version of Fallout 4 and 76's versions. Melee and unarmed combat is pretty much pointless because you cannot mod any melee weapon, or mod any spacesuit to allow for more unarmed punching damage. There's no Power Fist equivalent either. The guns feel fine to shoot though.
I've seen the phrase "a mile wide and as deep as a puddle" for Starfield discussion and it's 100% true. The story is passable, and I've been enjoying the faction quests and other sidequests I stumble upon(Operation Starseed in the Charybdis system for example), but once I finish the story, I'll probably shelf Starfield and go back to Skyrim or Fallout 4 if I want an open world Bethesda experience.
Also, who consciously decides not to make local city maps? It's easier to navigate them once you become familiar with them but it's still baffling that there's no proper city maps. :
Have to disagree on the side missions. Most are meaningless fetch quests. You get better side quests in MMOs...
The faction quests are the only missions worth playing. The main plot has exactly two missions that aren't fetch quests. I am baffled that the UC Vanguard plot is not main missions. What airhead at Bethesda thought that new game plus makes for a great main plot?!
@@haariger_wookie5646 That's fair about the side quests. A few stick out for me even if the objectives are basically fetch quests, like Operation Starseed. I agree with the rest though. Funny enough, I picked up No Man's Sky on a sale this past weekend so I'll be trying that out once I finish Starfield's story lol
Thing is bethesda sucks at most things. They suck at story telling, character writing, making interesting gameplay... Their main strength is world design. It's the one thing they've been consistantly getting better at, even with FO76, and with starfield they just threw that out of the window. Even the handcrafted places are impacted because they needed to tone down the designer's creativity in order for it to fit seamlessly with the procedurally generated parts.
If ressource farming was mandatory people would be saying it's "forced busywork"...
There is way more weapons variety in Starfield than Fallout 4...
The quest of Starfield are miles better than in previous Bethesda games, the writing and worldbuilding in general is very good, this game is pretty deep.
City maps would just make it so the player spend more time in the menu, it's more fun to learn the city yourself or to use the in game indications
@@haariger_wookie5646 That's just false, all of the side quest have some sort of narrative, a fetch quest would be a quest without any narrative.
It's like you never played RPG, in most RPG the side quest will involve going to pick up an item, clear a location or kill someone, but like I said what really matter is the narrative of the quest.
When I was a kid, my favorite show was ER. My all-time favorite episode of ER didn't take place in a hospital. The entire episode was shot in a small classroom with only 5 of the massive ensemble of more than 15 actors. It was great. The drama was intense, the emotion was high, the dialogue was biting, all top-notch. The intimate setting made that episode what it was.
I feel you. I'm still in act 1 of baldurs gate because i play it slowly, take my time, try to see as much as possible and just enjoy what was given to me.
StarLoader
4 times the polish.
16 times the space.
It just loads. - Godd Poward.
Its also like 125 gigs to download and takes up 140 GB of hard drive space. I have no idea how they managed to make it a bigger size than Red Dead 2 but they did. Devs should really start making 2k/4k textures an optional download (Like Diablo 4 has) cuz these file sizes are getting insane.
Honestly what I am really impressed by is the procedural generation in remnant 2. They wanted to make it much more focused around center points so that the game has a focused, enjoyable plot but has some variety when you replay the game which makes replays much more enjoyable
even there, the dungeons have only a few different layouts, most of the "randomness" is which dungeons you're going to have spawn in your playthrough. Its very good because most of the game is intentionally designed, there's just a bit of shuffling to keep things fresh.
Remnant 2 does it well because it's hand crafted tiles, that are then shuffled around within parameters. It's not truly procedurally generated like a game like No Mans Sky. After you play that game for a hundred hours or so, and constantly secret hunt. You start noticing patterns in the game and can hunt areas down based on landmarks.
Remnant is the only game in which I can't wait to see how they keep mastering their way of using Procedural Generation with subsequent releases.
Absolutely a banger of this year and to think we haven't even gotten to the first expansion but yet have such a kickass base game to play.
Remnant 2 is just a damn good game in many ways. One of the most fun experiences this year, and given how big this year has been, that's saying a lot.
It also helps that Remnant's fundamental gameplay loop is well-tuned and extremely entertaining - despite the procedural generation, there's a lot of GAME there, and as you say the procedural elements they include are curated to augment that rather than strictly being designed for scope. The problem becomes when a game (like Starfield) throws together a bunch of mechanisms and crosses their fingers that a game will somehow appear.
freeside in new vegas has 80 times more life & character than all of starfield
Even the city of Morrowind has more Soul.
It also.....has stuff were not even allowed to sell, people! Only at Mick and Ralph's!
85 hours in Starfield and it feels like a drag. I had 120 hours in Baldur's Gate 3 and it felt like it was only 30 hours. Baldur's Gate 3 is out of this world amazing.
Why the fuck would you play a game you don't enjoy for that much time? Like I hate Starfield too, but at least try to move onto something else.
Why are you forcing yourself to suffer with Starfield?! 😂
Why play something that is a drag for 85 hours
@@henryjohnson-ville3834because deep inside down he believes that something interesting is gonna happen. Also since you invested so many hours in this game you want to finish it for sure.
@@revbenSame happened to me with Skyrim. My bare maximum is somewhat around 50+ lvls, after that I got so bored that just deleted the character. The game itself is just a clicker without any gameplay itself, without story, graphics or mechanics, but for some strange reason I was thinking that this is going to change somehow. That was 2 years ago. Well, the main logic behind this torture is just an idea that you will see new enemies, new locations, new something. I felt sorry for spending hours into this, like really sorry. But sooner or later I had to admit that's not going to happen, it's just empty and never ment to be filled with anything. So I think that's the reason why people play "drag" games. Coz they "invested" so many hours that they can't admit it was for nothing.
The first games I got into were traditional roguelikes, and my opinion based on that experience is that procedural generation only works if you fulfill a few requirements:
- Handcrafted (or very constrained use of generation) setpieces.
- Mechanics focused on mastery.
- Procedural Generation that can facilitate those mechanics and setpieces.
Brogue has some of the coolest level generation that I've ever seen.
Yeah, my dad's been hooked on Spelunky for years - he's not really a speed runner type but the depth of mechanics mastery and understanding combined with unique situations where the generation puts problems next to each other makes for his crack (I'm only not a fan because I can't platform to save my life, especially in 2D)
I think the game's strategy of having handcrafted 'rooms' that get smudged, shifted, rearranged, and randomly altered gives a good balance between the maps feeling unique and the maps feeling intentional
I immediately thought of Brogue as well, also because the dev had a talk about proc gen in Brogue (it's on youtube), where he talked about developing proc gen being a skill as well as designing levels by hand is. The proc gen behind Brogue is a craft, the reason behind the huge(lava)lakes and chasms for example is because being able to see a bit of what you'll go to later on makes exploring fun. Seeing rooms on the other side of the lake is similar in that sense to seeing a tower in the distance in an open world game.
@@madeliner1682 You and your dad might also like "Unexplored" (the first one) then! From what I heard it has a good proc gen system made to make exploring fun (inspired by the way the dev's kid liked to run around the room when playing), and it has a similar focus on problemsolving and understanding mechanics. Also, it's top down, not a platformer.
I don't think I've seen that talk, I'll have to find it! @@smaug131
I'm 37, full time job, married, make a decent living.
Just finished the Mass Effect trilogy again, and enjoyed every minute of it. Started Andromeda, and in 18h decided it's not worth my time.
I'll gladly dump 1000 hours into something like Monster Hunter World. I refuse to waste time slogging through a mediocre game like Andromeda, because "mid" is not worth my limited time.
A couple months before Starfield came out I was talking with a friend about it. We'd both loved Skyrim and the Bethesda formula when we were in high school, but we both have Playstations so he was bummed about not being able to play Starfield. I said I wasn't really hyped for it because I expected it to be repetitive, way too big for my schedule, and full of filler content, even more so than Skyrim was (and I couldn't play that either if it came out today). He didn't believe me and thought I'm just trying to convince myself I'm fine with not being able to play it.
Everything I've seen of the game just convinced me of that first impression...
Your point about non-gamers being impressed by buzzwords also hits a note. You see it with VR and the metaverse, with techbros pushing how immersive web3.0 will be because it's big and realistic and you have to actually go from place to place physically in the digital world etc. and like, those are inconveniences in the real world that I'd be happy to do away with in the digital one... It's what made RDR2 not fun for me, when instead of a store menu UI they make you physically walk to the product you want to look at. Or like early smartphone UI conventions going for realism and retaining the inconvenience of what they're simulating, before realizing the potential of digital space to be easier to interact with.
Anyway sorry for the rant, I work in tech and am surrounded by non-gamer tech enthusiasts who don't know we've actually learned many lessons that go against how they imagine web3.0.
Gothic 1 is a 2001 RPG where creators had to come up with an idea onto why the location is small and limited. They made it into a magical sphere prison and suddenly technical limitation became a narrative feature
It also didn’t have loading screens. You could go anywhere, anytime, every NPC had a routine.
The only loading screens happen when moving to the next story chapter, at which point parts of the world change.
@@ArDeeMee , well, in Gothic I there were a few separate locations behind the loading screens... Old and New mines, a couple of temples. Same in the Gothic II. Still, the Colony and Khorinis was (is) a living-breathing experience you can immerse yourself into. One of my favorite RPG's! Among Gothic series, I think part III has a truly loadscreen-free world. But I love parts I and II more.
@@leh2012 Yeah, the few loading screens happened every time the world changed permanently, e.g. new chapter. In G2 it happened at those points, and if you went to the colony, aka an entirely different map. It’s the first game I ever saw where each and every NPC has a daily life. It was sooo good.
@@ArDeeMee , I was amazed by those game too... A good momories! But I think it can surprise even today those who missed it beack in the day (or those who are younger than the game). If they manage to run it on their modern PC's that is. Also, there are lots of mods available for the game some of which are basically a new games...
@@leh2012 Yeah, the worldofgothic website is a must. Widescreen, higher textures, etc. ^^
I think the procedural generation in Minecraft was and still is impressive because its the main soul of the game. Probably the only game I can think of where it is essential and makes the game what it is.
For something that's essentially the opposite of Minecraft, look at .kkreiger from 2004. It's a single-level FPS with graphics on par with Doom 3 that fits into 96KB. The level layout isn't procedural, but everything else is and is built from instructions at load-time. Not that interesting as a game, but definitely a technical marvel.
Josh's point about limitations breeding creativity is actually a big part of why Artix games are so great (some of them, I mean). One of the developers of Dragonfable was talking about this exact matter only a few days ago in the Falcon's Nest discord server, they're forced to get creative with character models due to the limited number of animators they have.
(PS We at FN are eagerly waiting for the DF video, Josh!
Doctor Who's penultimate season 4 episode "Midnight" was written because they had to replace an episode that wouldn't be done on time, so all they had was a bus set. And they made one of the best episodes ever. With no props, little CGI, and no money. All excellent writing and using limits.
You nailed it once again Josh. Procedural generation is good for certain games but not all. BG3 is an example of a long RPG I want to make time to play due to how compelling it is and how when you explore you find quality side content and interactions. Starfield has no appeal to me. I do enjoy No Man's Sky but it's a game I play on and off when I want to wander and chill out.
I would like to say that an example of where procedural generated content works in an RPG is Wildermyth. Each world in a campaign is made from scratch, your characters can be as customized or random as you'd like, and beyond the overarching campaign storyline, everything that happens in between is randomly determined as you go. Obviously it's not 100% random, but each campaign promises to be a unique experience.
Yesss. Honestly Wildermyth is an incredible game, the writing especially really helps to pull the whole world together.
What I hated the most in Starfield is not the emptiness but the missions are extremely linear. Even when the game gives me different dialogs I realize no matter which dialog I chose the result was exactly the same. Ex: we had to deal for an artifact and there was a dialog option to attack the man I was completely deny to attack him. No matter which option I chose the end result was exactly the same. I only had the illusion of choice. There is the pirate mission where my partner wanted me to side with him to take out the leader. In the end it made no difference, the end result were exactly the same as if it didn't happen.
When everything in the game is up in the air, how can the missions be anything else but linear? Think about it - to implement a story mission, you have to know the locale, the layout, everything about the area. When it's all a diceroll, you cannot reliably account for anything that the dice will affect - so you can only write absolute barebones generalities and HOPE the dice roll some cool stuff. Sometimes they do, sometimes they don't - but most of the time, the dice roll their mathematical average.
The artifact quest really killed the game for me, I know Bethesda well enough and I didn't expect much freedom of choices, but that quest showed how thoughtless the quest structure is. I talked to the guards and got some hint for "hidden path", then I managed to persuade the collector to show me the artifact, I thought I could pretend to leave first and sneak back since I knew the hidden path, then I realized that the collector just stay at the vault and never leave, I back stabbed him and found out that he cannot be killed instantly and will always trigger the dialog of "take the artifact and get the bounty" or "no witness but get the bounty anyway". I loaded the save and pick pocketed him to get the key, then sneaked to the artifact, "bounty added". I reloaded again and killed everyone on the ship and took the artifact, "bounty added". I know the bounty was meant to trigger the Crimson Fleet quest, but the choices make no sense, why the quest ask you to "steal from the thief" and "do it peacefully if possible", and why was I getting the bounty when there were no witness or even no one alive.
@@thesunthrone no, lack of choice even appears at the end of unique side missions, which are irrelevant to the main story. The OP has a point. I don't mind most of the flaws, but that's just lazy. Far too many characters are immortal - perhaps they need to be used for DLC, but even so, that's lazy/cheap too. Anything that's worse than Skyrim/Fallout in Starfield is bizarre. I guess they started off too ambitious, broke the engine, so had to remove stuff - a bit like GTAIV after GTASA, but not as extreme (and not a different engine). Most of Starfield is typical Bethesda - people were expecting a brand new game? Naive - it was always going to just be a Fallout reskin. It's a good game, Brent - if you like Fallout/Skyrim, albeit nowhere near Skyrim's level, although combat's more fun. Typically addictive and atmospheric, but typically technically poor too (compared to GTAV, RDR2, and Halo Infinite, which should be expected).
@@goatlps Don't mistake me for excusing Starfield - I'm just saying what procedural content means. When you roll two six sided dice, 50% of the rolls will add up to an average of seven, that's just statistics. But what good are quests that can work only 50% of the time?
The procedural design necessitates that quests IGNORE the procedural design entirely for them to even work - and because of that they'll just be the most blandest, most boring fetch quests imaginable, because nothing else can even work.
@@thesunthrone makes no sense. You can create procedural environments and interiors and still have good quests. They just didnt bother making good quests. Except a few ones, like the one in a lab where there's a dude stuck in some kind of other dimension and you have to get through portals to naviguate between both realities. Environements are procedurals, quest is interesting enough.
When I woke up today, I did not expect to find out that Josh Strife is actually just Matt Berry.
Procedural Mirrors Edge would be awesome! One of the reasons I enjoy speedrunning missions in Warframe. Great movement/parkour through constantly changing levels.
Procedural generations works in Warframe because of the gameplay loop.
But, DE didn't just throw some random prompt. They created each object and room by hand according to the theme of the planet or level and then let the game decided which room connect to which.
Yes, that occasionally would create a nonsensical level design like a 2 or rarely 3 or in very rare case, 4 identical room created one after another and weird dead ends that logically shouldn't built like that in a space ship. But we just accept it is as what it is because tbf we don't gaze the scenery. We grind til morning for materials.
The Deadpool fight scene situation is so perfect because it's completely within character for him. They managed to make a limitation work with an established character. That's great.
Thats one of the things I really love about Breath of the Wild and Tears of the Kingdom. The map is just packed full of little curated details. It's amazing.
Yup, and despite being on the Switch’s archaic AF hardware, the world is more DETAILED than 95% of open-world games (and I DON’T MEAN in terms of visual fidelity because it’s easier to make something look nice and flashy as opposed to making sure the “world simulation elements” can respond to any number of things you can do).
For BOTW I can agree. Totk though, even though the sky islands are fairly good, the underground is pretty awful and barren. This is made especially obvious when you start lighting it all up with the light roots
@fuzzydude64 I actually really like the underground 😅😅
To each their own though
@@fuzzydude64
Well, you can KINDA say the same thing about the shrines in BOTW where some had FAIRLY BIG SPACES within them but not a whole lot to interact with outside of whatever elements pertaining to the puzzle you needed to solve there.
Level 1: Hay Bale
Level 2: Hay Bale + Tractor
Level 3: Hay Bale + Tractor + Building
Level 4: Tractor
Level 5: Tractor + Building
Level 6: Tractor + Building + Hay Bale
Level 7: Building
...😂
To be fair, the scale of games used to be impressive. Wow was mind blowing because of how big the world was and how much time you could spend on exploring it through different factions/classes. Now, we are more conscious of the padding and the repetitivity of these games.
But WoW was hand crafted and even the vanilla world still feels pretty cool.
Hi-Fi Rush is a perfect example of this. Hi-Fi Rush will take about 8 or so hours to complete, but the amazing gameplay, music, and visuals will make you want to play more. In my case, I went and spent 82 hours getting all the achievements in Hi-Fi Rush. I wouldn't have done that if Hi-Fi Rush wasn't as good as it was. I chose to play Hi-Fi Rush for 82 hours, the game didn't expect me to slog through 82 hours if subpar gameplay and visuals to get to the end of the game.
As for budget limitations; I think it was DS9 where they had to scrimp and scrape the first season and one of the best, yet cheapest episodes from that was Duet. Brilliantly acted and written with no grand battles or shoot outs or anything not even phaser or disruptor fire. Just good solid dialogue.
You're not wrong about the game-length aspect.
BG3 for example by the time I got to Act 3 I was running to the finish line. I'd experienced everything gameplay wise it had to offer, most companion stories had wrapped up - it was just the story left so I didn't even want to explore any side content just reach the end already.
Agreed completely. The only game with a narrative that I've played where I enjoyed the procedural generation was XCOM 2/WotC. Its absolutely amazing because it forces you to adapt your tactics, but they made sure geometry doesn't break any more than it did in XCOM:EU, and it is a strategy game. I can't say the game would have been better with set maps because it would come across the issue of every mission being the exact same, and I honestly wish more tactical games were similar.
Yes. I agree the OG XCom games are a prime example of good procedural generation utilization.
So many half-thoughts i didn't know how to express finally put in a coherent, eloquent format. This channel is a treasure.
I think one of the other big problems is that the Procedural Generation engines people make tend to be way too simple. Like you said, start in a forest, there's an object you'd see on a farm, there's another one you see on a farm. The engine didn't make significantly different ones.
Too simple' does not reflect the essence of the matter - AI will not write a good book for humans. It can only make monsters out of pieces 'of other people's work, but he won't tell the story because he doesn't feel it or understand it. Moments full of emotions in games (or music or movies) are a symphony of emotions and means of expression (dialogues/words, sound and images) - AI is not able to recreate it.
@@mwmv921 yeah but not every moment in a game needs to be a big deal dramatic climax - there a lot of value in allowing less plot relevant experiences that let the player understand the themes and mechanics of the game. Those sections can be procedurally generated, if you do it right. But to do so, you need to take the time to design the hell out of those systems, making sure it produces sufficiently different but related experiences every time. An entire game cannot be made by a computer, but I think there are a lot of segments that can be, if the devs took the time to do it right
Early No Man's Sky was a lot like that. Thankfully, I bought it used a year after release and not at release, but after about 30 hours in the grind became boring. I still revisit it when there's updates, but exploration doesn't have any appeal to me in NMS anymore. Contrast with Minecraft, where exploring procedural caves still feels interesting because of the challenge in it (especially in early game)
Imagine a space exploration game with the same budget, but instead its based in one star system and just has four or five planets, each of which has a rich variety of biomes, factions, locations, characters etc on it to explore, plus some orbiting stations and asteroids and stuff. That would be so much more immersive to me than any million procedural planets in this or no mans sky
Play Outer Wilds. It is pretty much that.
you just described Outer Wilds to a T. It is, hands down, my favorite game of all time.
Agreed, it's my favourite game aswell.@@Palakarava
I mean, it could be two or three systems even, with 20 or so curated planets with a hypergate or some such in between. Could still have the scale but more of the quality that makes a good game.
Starfield has a rich variety of biomes, factions, locations and character to explore. Starfield also has orbiting stations and asteroids.
All of the content in Starfield is handcrafted, the procedural generation is only for the landscape.