So with 8 players and 41 civs it comes to 95,548,245 possible combinations. With one game every week T90 can only do this another 1.8 million years. With around 65 years remaining he will cover 0.0035% of all possible combinations. The takeaway is: T90 can't stop at 95 years old
But what if the same combination occurs twice? Well If T90 finishes 1 game every hour and we assume that he plays 10xcivbonus 24/7/365 for the next 65 years. Then the probability of a combination of the same civs occuring twice is at 0.000017%.
41:40 I used my stopwatch and some math to see how close it was. From 9.0 years to 1.0 years took 23.71s, that puts the rate at 2.96375 seconds per "year". So the estimated time from 9.0 to 0.0 would be 9*2.96375 which is 26.67375 seconds. The readings I took for 9.0 to capture was 26.66 seconds. Based on all of that the time they had left over was 26.67375s - 26.66s, which is 0.01375 seconds. TLDR: 0.01375 seconds to spare. Very very close. This result is not super accurate, the margin of error of my stopwatch is larger than the value we are looking for. I'm sure a frame by frame analysis would yield better results. If you are wondering why I am measuring from 9.0 to 1.0 to get the rate, I had tried gathering across a larger span (20years) of the year countdown and I got a result of 2.962 seconds per year, a result I am very confident in. It appears that when the game displays a tenths place the rate of seconds per year changes slightly. I can't really tell with just a stopwatch.
Good job! Great math! I also think 13ms is irrealistic. The calculated 2.96s per year looks good because you spread the measuring error with the division but for the second measurement (26.66s) the measuring error is TOO RELEVANT since you are just going to subtract it from the expected time (26.67s). Measuring over a long period won't help us there. You could measure many times and average. they'd all be around 26.6s. you could get 26.4s (meaning we had 200ms left) or you could get 26.8s (which doesn't make any sense if the game was expected to finish at 26.67s). Anything can happen, getting 13ms from the expected time was a coincidence TL;DR: the answer can't be 13ms if you have a measuring error ~200ms that is your reacting time *I guess we could say there was less than 200ms left, which is still impressive* Ps. I got consistently 150ms reacting time in a game on my phone but it's difficult. You are slightly distracted and it's over 200ms. Ps2. I wish the game would make 1y=3s.
I loved your enthusiasm because I'd do the same. Ps3. Trust me I'm an engineer :) Back at uni there was this physics lab where we had to measure things then make calculations while considering how the error is spread. Lots of formulas. Pretty cool, give you some insights on how error is nasty. I say this because I saw you measure time with 2 meaningful digits then multiply with more than 2. Those extra digits won't help you, despite best intentions. I don't remember if you could use statistics to mitigate measuring error: like measuring a bazillion times, then using some cleaver math and improve your results. Damn statistics is hard (and useful). At first glance just averaging a normal distribution could indeed help.
@@andrevc85 Yeah I have no idea how to calculate error. I know it is there, but that part of statistics went over my head. I had a feeling the extra digits would not be useful. (If you know of a good explanation showing in depth the how, why, and when that is important, that would be appreciated.) ...but my goal was to play it safe and keep the data lossless. I have a habit of not rounding my floating point numbers. In simulations if you iterate a float, rounding every iteration, that will make any error increase and propagate. That has nothing to do with what we are talking about, but that was just my frame of mind. Yeah, I thought about calculating the upper and lower bounds for all of the variables over the coarse of several measurements, to kind of see the bounds of the whole thing, but I got lazy and decided to leave that up to the viewer. :) It also became obvious rather quickly that reaction time was bigger than the answer. I thought about trying to eliminate that by sampling multiple times and picking a value in the middle but that would require a bit more attention. Ideally if there was a way to have a debug version of the aoe playback software, something that would output the low water mark of the countdown, that would tell you how much time the simulation believed was left. From there it would be easy.
Update: I downloaded the video and put it in blender to step through each frame. The results are not very satisfactory. At a rate of 2.96 seconds per year I would expect to get 177.60 frames per year given 60 frames per second. Instead I get 88.50 frames per year. The only thing I can think of that might be causing problems is that there are likely a sizable number of dropped frames. It is very possible that youtube has done some compression that is making stuff weird. It is possible that ffmpeg messed with stuff. Maybe the video on t90Official's harddrive is better. Maybe his recording setup dropped some frames. The number of frames from 1.0 to capture was 90. If I pick the "best numbers" for frames per year, (instead of using 88.50), I get 94-90, 4 frames. 1/60 * 4 = 0.066 seconds (6 repeating, yes I know too much digits.) If we average my previous calculation with this one we get (0.01375 + 0.06666)/2 = 0.04021 seconds (Very hand wavy) But none of this data is trustworthy because of all the dropped frames. Still, it was fun stepping through the game frames. The fire, the health bars, the units, particle effects, all change the sprite they render on different frames. It is almost like the game does not fully re-render everything in a single frame. It makes sense, considering this game was designed to run on computers with megabytes of ram.
As a regular 10x player I want to notice that most versions have not fixed the saracens market, but have the return very low. They usually cap it that sell & buy has the same number. But when you sell, the price gets down so after buying again, you get 2gold net return for 100res sell/buy or 10 if you do 500res sell/buy. Combining this with hotkeys that you just need to keep pressed makes it still easy to get a lot of res on basically any 10x mod
This may be the best game I have ever watched. Incredible. Teal's ram train, the shitbows, the scorpion army, the invincible castles, the last minute rushes of middle.... Just incredible.
So persians in 10x shared bonus mod has to research a free tech to activate the TC HP. The reason for this is because if there is a byzantine ally they will get too much HP and no longer be able to have a TC. As for why this tech also removed the negative food and gave them negative gold instead, I actually have no idea.
I think it really is the Siege Ram's splash damage. It works differently from other forms of Splash Damage and no one would ever think to test that; you can't place buildings next to the monument and no one normally uses rams on units.
but if you slow the video down, there are moments where the rams didnt attack anything and the monument still took damage, the only thing hitting the monument in those instances were the scorpions, though it may be only specific hits that count
That 0.1 was SO HYPE. And then when you went and showed Reds Rams blocking that choke point I was like "omg, its thermopylae" And then I had the realization...RED IS PERSIANS! IT REALLY IS! THOSE RAMS ARE SPARTANS! HAROOM! HAROOM! HAROOM! LETS GOOOOOOOOOOOOO!
It's been way too long since the last time! Great so see that UU are no longer busted via their elite upgrade. Still waiting for the day we are blessed with Goths for 300 pop Edit: I didn't even think of their free infantry and insanely fast production speed on the baracks. 200% from their team bonus and 1000% from perfusion?
That was insanely close... well played Bennu... extremely unlucky in first instance, extremely lucky second time around... Hats off to him to win this crazy game! And hats off to all others too for AMAZING content and loads of fun! Much better than to watch hollywood movies :)
EVERY SINGLE DAMN TIME someone claims the monument in the middle and starts the victory countdown timer before everyone has researched all their techs and before anyone has changed their diplomacy settings to enemy all. You need to ask if everyone is ready, get all 7 "yes" or"1" and THEN claim the monument. T90 if you could quiz people on this procedure and see if they understand it before they're allowed in the game because someone gets this wrong EVERY SINGLE TIME!
I've only ever been a very casual aoe2 player, but recently found the more competitive/serious side of it and I love how weird some of the games are. Imo it's a sign of a good community
Something that I don't think was mentioned was cheaper hussars. I'm pretty sure Magyar light cav line is discounted, so theoretically they would have been just plain free. But without any other bonuses they were just getting melted by scorpions and pretty much anything else anyone was making so they didn't do all that much.
This game really lived up to the hype. Normally I watch your AOE videos on 1.5 speed because, well, let's face it, I'm used to Starcraft 2 speeds for things actually occuring in a game now, but eventually I realized that was totally unnecessary here. The normal speed looked like it was like, 2 or 3 speed. Completely ridiculous meme units, but also really gripping and thrilling close gameplay, someone taunting only to get karma for their hubris, rams moving like a freight train through whole cities... great time.
I've done this matchup many times, psa RAMS COUNTER SCORPIONS in these settings. Not only do rams kill scorps quite well themselves, but with 1.9k HP and only taking 1 damage a hit, it's cheap to counter them
@t90 what would the burgondian flemish revolution do in this game turning your eco into spier men that are stronger then halbedier( atleast wen i last played) but would it do it for all players, and will it take over all the buffs from the villager or not
The instructions are so clear: get all unique techs, then enemy and go for the monument. I don't understand why people have so much trouble with these simple instructions.
So with 8 players and 41 civs it comes to 95,548,245 possible combinations. With one game every week T90 can only do this another 1.8 million years. With around 65 years remaining he will cover 0.0035% of all possible combinations.
The takeaway is: T90 can't stop at 95 years old
That's without repeating and the game doesn't account for it so the real number it's bigger
But what if the same combination occurs twice?
Well If T90 finishes 1 game every hour and we assume that he plays 10xcivbonus 24/7/365 for the next 65 years.
Then the probability of a combination of the same civs occuring twice is at 0.000017%.
What are the numbers with 45 civs?
With 45 civs there are 215,553,195 combinations, so it'll take more than twice as long
@@daanwilmer Nice.
I like that Forest Nothing is such a normalized map on T90's channel that it's not mentioned anywhere in the description or title.
41:36 look at the rams blocking off the elephants in the top right corner, amazing
Teal is one guy that draws in new people for these kind of shenanigans
47:40 Scorpions being chased by rams felt like FAST AND FURIOUS AOE2!
At this speed, is it technically possible for an onager to kill itself, by running too quickly into its own flying shot? :D
Time for AoE mythbusters.
We need Spirit of the Law :D
41:40
I used my stopwatch and some math to see how close it was.
From 9.0 years to 1.0 years took 23.71s, that puts the rate at 2.96375 seconds per "year".
So the estimated time from 9.0 to 0.0 would be 9*2.96375 which is 26.67375 seconds.
The readings I took for 9.0 to capture was 26.66 seconds.
Based on all of that the time they had left over was 26.67375s - 26.66s, which is 0.01375 seconds.
TLDR: 0.01375 seconds to spare. Very very close.
This result is not super accurate, the margin of error of my stopwatch is larger than the value we are looking for. I'm sure a frame by frame analysis would yield better results.
If you are wondering why I am measuring from 9.0 to 1.0 to get the rate, I had tried gathering across a larger span (20years) of the year countdown and I got a result of 2.962 seconds per year, a result I am very confident in. It appears that when the game displays a tenths place the rate of seconds per year changes slightly. I can't really tell with just a stopwatch.
Good job! Great math!
I also think 13ms is irrealistic. The calculated 2.96s per year looks good because you spread the measuring error with the division but for the second measurement (26.66s) the measuring error is TOO RELEVANT since you are just going to subtract it from the expected time (26.67s).
Measuring over a long period won't help us there.
You could measure many times and average. they'd all be around 26.6s. you could get 26.4s (meaning we had 200ms left) or you could get 26.8s (which doesn't make any sense if the game was expected to finish at 26.67s). Anything can happen, getting 13ms from the expected time was a coincidence
TL;DR: the answer can't be 13ms if you have a measuring error ~200ms that is your reacting time
*I guess we could say there was less than 200ms left, which is still impressive*
Ps. I got consistently 150ms reacting time in a game on my phone but it's difficult. You are slightly distracted and it's over 200ms.
Ps2. I wish the game would make 1y=3s.
I loved your enthusiasm because I'd do the same.
Ps3. Trust me I'm an engineer :)
Back at uni there was this physics lab where we had to measure things then make calculations while considering how the error is spread. Lots of formulas. Pretty cool, give you some insights on how error is nasty.
I say this because I saw you measure time with 2 meaningful digits then multiply with more than 2. Those extra digits won't help you, despite best intentions.
I don't remember if you could use statistics to mitigate measuring error: like measuring a bazillion times, then using some cleaver math and improve your results. Damn statistics is hard (and useful).
At first glance just averaging a normal distribution could indeed help.
@@andrevc85
Yeah I have no idea how to calculate error. I know it is there, but that part of statistics went over my head. I had a feeling the extra digits would not be useful. (If you know of a good explanation showing in depth the how, why, and when that is important, that would be appreciated.) ...but my goal was to play it safe and keep the data lossless. I have a habit of not rounding my floating point numbers. In simulations if you iterate a float, rounding every iteration, that will make any error increase and propagate. That has nothing to do with what we are talking about, but that was just my frame of mind.
Yeah, I thought about calculating the upper and lower bounds for all of the variables over the coarse of several measurements, to kind of see the bounds of the whole thing, but I got lazy and decided to leave that up to the viewer. :) It also became obvious rather quickly that reaction time was bigger than the answer. I thought about trying to eliminate that by sampling multiple times and picking a value in the middle but that would require a bit more attention.
Ideally if there was a way to have a debug version of the aoe playback software, something that would output the low water mark of the countdown, that would tell you how much time the simulation believed was left. From there it would be easy.
Update: I downloaded the video and put it in blender to step through each frame. The results are not very satisfactory. At a rate of 2.96 seconds per year I would expect to get 177.60 frames per year given 60 frames per second. Instead I get 88.50 frames per year. The only thing I can think of that might be causing problems is that there are likely a sizable number of dropped frames. It is very possible that youtube has done some compression that is making stuff weird. It is possible that ffmpeg messed with stuff. Maybe the video on t90Official's harddrive is better. Maybe his recording setup dropped some frames.
The number of frames from 1.0 to capture was 90. If I pick the "best numbers" for frames per year, (instead of using 88.50), I get 94-90, 4 frames. 1/60 * 4 = 0.066 seconds (6 repeating, yes I know too much digits.)
If we average my previous calculation with this one we get (0.01375 + 0.06666)/2 = 0.04021 seconds (Very hand wavy)
But none of this data is trustworthy because of all the dropped frames. Still, it was fun stepping through the game frames. The fire, the health bars, the units, particle effects, all change the sprite they render on different frames. It is almost like the game does not fully re-render everything in a single frame. It makes sense, considering this game was designed to run on computers with megabytes of ram.
Love these games, you need to add more of these 10x tech games, so much fun and entertaining.
As a regular 10x player I want to notice that most versions have not fixed the saracens market, but have the return very low. They usually cap it that sell & buy has the same number. But when you sell, the price gets down so after buying again, you get 2gold net return for 100res sell/buy or 10 if you do 500res sell/buy. Combining this with hotkeys that you just need to keep pressed makes it still easy to get a lot of res on basically any 10x mod
42:05 absolutely legendary moment.
This may be the best game I have ever watched. Incredible. Teal's ram train, the shitbows, the scorpion army, the invincible castles, the last minute rushes of middle.... Just incredible.
So persians in 10x shared bonus mod has to research a free tech to activate the TC HP. The reason for this is because if there is a byzantine ally they will get too much HP and no longer be able to have a TC. As for why this tech also removed the negative food and gave them negative gold instead, I actually have no idea.
These are some of my favorite videos/commentaries. The hilarity of it all makes it so funny.
It was literally at 0.1 years. What. The. Hell.
I think it really is the Siege Ram's splash damage. It works differently from other forms of Splash Damage and no one would ever think to test that; you can't place buildings next to the monument and no one normally uses rams on units.
but if you slow the video down, there are moments where the rams didnt attack anything and the monument still took damage, the only thing hitting the monument in those instances were the scorpions, though it may be only specific hits that count
That thumbnail is one of the more insane ones you have made so far. What a good game for the new year!
That 0.1 was SO HYPE. And then when you went and showed Reds Rams blocking that choke point I was like "omg, its thermopylae"
And then I had the realization...RED IS PERSIANS! IT REALLY IS! THOSE RAMS ARE SPARTANS! HAROOM! HAROOM! HAROOM! LETS GOOOOOOOOOOOOO!
That was insane. Love these custom games, and T90, your commentary makes it that much more fun man. Thank you!
49:03 Having a friend sing Bohemian Rhapsody does sound like a good time but it's very much an "odd flex but ok" moment
39:00 Like watching and listening to T-90 narrate a game about his ant farm.
It's been way too long since the last time! Great so see that UU are no longer busted via their elite upgrade. Still waiting for the day we are blessed with Goths for 300 pop
Edit: I didn't even think of their free infantry and insanely fast production speed on the baracks. 200% from their team bonus and 1000% from perfusion?
42:00 in a game that's all about strategy, macro, micro etc did I ever think we needed slomo replay to find out what happened 11
38:30 We wore Tommy Hilfiger and North Face back in the late 90s, my man.
dude this might be a top 3 game i've seen. KoTH already has the effect of making good nail-biters, but holy crap. twice!!
I love these 10x Videos so much, such a joy to watch the chaos unfold.
watching the rams run around buildings like a benny hill sketch, lol
That was insanely close... well played Bennu... extremely unlucky in first instance, extremely lucky second time around... Hats off to him to win this crazy game! And hats off to all others too for AMAZING content and loads of fun! Much better than to watch hollywood movies :)
I love how yellow has a casual 4 vills complete a castle while 8 siege rams on it.
EVERY SINGLE DAMN TIME someone claims the monument in the middle and starts the victory countdown timer before everyone has researched all their techs and before anyone has changed their diplomacy settings to enemy all. You need to ask if everyone is ready, get all 7 "yes" or"1" and THEN claim the monument. T90 if you could quiz people on this procedure and see if they understand it before they're allowed in the game because someone gets this wrong EVERY SINGLE TIME!
Chill
@@ozziey52Don't act like every single person watching these isn't a wierdo
@@-BuddyGuyI’m not
@@bridgenorton537 Embrace it
Yeah it drives me nuts too lol
When teal was zooming around with rams I was just like "An army of rams? How quaint"
This is great, but I would like to see what 10x bonus could do on water too!
I've only ever been a very casual aoe2 player, but recently found the more competitive/serious side of it and I love how weird some of the games are. Imo it's a sign of a good community
I’d love to see this game 5v5 random, but I just don’t want the game to end with the trade possibilities
Random+ Mirror civs solves that problem, but game could be predictable then.
Yellow at the end of the game:
"Remember, the enemy's gate is down."
"not a Daut castle" 🤣🤣🤣 I'm pissing myself
My god what a close game. Had me shouting on the edge of my seat
That reset on 0.1 years was INSANE!!!
That was so fun to watch. Keep these coming :D
Tired as hell but this game was so exciting that it woke me right up. Best 10x shared mod game in a while, perhaps the best yet!
T90 That was brilliant to change the color on the thumbnail to teal. What a great attention to detail for YT viewers. Well done, sir.
The title should have been... Benny and the Jets. 🤣
This is Age of Modern Warfare with those Scorpions :)
Something that I don't think was mentioned was cheaper hussars. I'm pretty sure Magyar light cav line is discounted, so theoretically they would have been just plain free. But without any other bonuses they were just getting melted by scorpions and pretty much anything else anyone was making so they didn't do all that much.
41:50 *0.1 YEAR LEFT!* INSANE 11
Imagine Incas Poles and Spanish. Those Vils would be unbeatable
8 players in a 10 times tech battle PLUS big trees? I'm in!!!
“Choo choo mother effer, let’s go!”
wowowoow what a moment!!! 41:30
Will you ever do a x10 NON-Shared game? Like on a regular map.
I thought the thumbnail has GOT to be clickbait... but they delivered. Absolute legends.
Broooooo, like 0,5 seconds to glory... What a great game ❤❤
I literally just watched one of the previous 10x shared civs like 3 hours ago.
This game really lived up to the hype. Normally I watch your AOE videos on 1.5 speed because, well, let's face it, I'm used to Starcraft 2 speeds for things actually occuring in a game now, but eventually I realized that was totally unnecessary here. The normal speed looked like it was like, 2 or 3 speed. Completely ridiculous meme units, but also really gripping and thrilling close gameplay, someone taunting only to get karma for their hubris, rams moving like a freight train through whole cities... great time.
Blue: "Kept you waiting, huh"
This is not even clickbait. Incredible
BENNY is 100% the true winner.
I'm imagining teams 10x tech where they have Caravanserei. Trade carts going down the road like Formula 1 racers.
Imagine Bohemians in this. With all those siege workshop bonuses.
12:00 *PTSD x256 tech tree nothing Spanish Civ
I also want to be 95yo and in a nursing home and having the staff tell me to turn off the computer and stop listening to your casts haha
Yellow managed to do the unthinkable with his -951 gold at the biginning. He was in depth in Ago of Empires. #IsAOE2BecomingSimcity
Red with the siege ram wall to block the stream of elephants was an inspired play.
You're the best T90! This was too much fun 😂
Mangos show up
"let's get ready to rubble"
I've done this matchup many times, psa RAMS COUNTER SCORPIONS in these settings. Not only do rams kill scorps quite well themselves, but with 1.9k HP and only taking 1 damage a hit, it's cheap to counter them
We appreciate you t90! Keep that energy no matter who's around lol
Game was nuts, but can we talk about how everyone let yellow fortify the middle after the first bit of craziness?!
For somebody who plays this game for a living, I’m amazed at how little T90 knows about the tech trees 😂
Blue and teal were MVP for entertainment in this game
This is awesome... can't believe what my favourite game has morphed into
I think this was the best game I’ve ever seen on your channel
that was one of the most intense game i've ever seen.
I was thinking someone would wall his elephants first time he almost won, but nope, turns out mobile tank rams are much better :P
Man, that was fantastic. What a game!
Finally a scenareo where fortifications do what they are build for 😎
Hey @t90 I made a mod that have all UTs researched in dark age. This way there's no more coordination hassle!
Bump
Near the end Blue was just rampaging through Yellows base. Wonder what they did you to that. Looked like a blue blob vacuum cleaner.
Spanish build speed x10 is hilarious to see, it looks like they have "aegis" on 11
Finally 10x Tech again, I love it.
Best x10 techs game i've seen in my life, thanks T90
@t90 what would the burgondian flemish revolution do in this game turning your eco into spier men that are stronger then halbedier( atleast wen i last played) but would it do it for all players, and will it take over all the buffs from the villager or not
I think that was the best game I've ever seen.
Please update us on how close that timer was!
i always find the 1111 for laughing to be a bit confusing, i always imagine 1 as yes, 2 as no, and 5 as laughing
Imagine a Teuton with a persian with Citadel
Burgond + spanish + turcs for the ultimate BBC
10 x shared civ bonus plus random cost mod lmao! Let's try it T90.
I watched an old vid of yours and the 10X civ bonus gave the vils super power but they also had the civil that could turn all the vills into infantry?
The instructions are so clear: get all unique techs, then enemy and go for the monument. I don't understand why people have so much trouble with these simple instructions.
I would have felt bad if Yellow lost it again, but I would also have laughed.
Wish we had Romans for the scorpions and Turks to improve the bombard cannons. Excellent game nevertheless 👌👌👌
what a crazy game! red is MVP and yellow deserve the win
that woman definetly is a keeper
Yellow: What a legend
Yellow villager, Wendy the Welder of AoE.
HOLY CRAP a 0.1 stop! 😂😅
Elephants, Rams, Scorpions, this is doctor dolittles game
What an insane game! Thank you!
That was ridiculous but oh so much fun lol
I support the 95 year old retirement home casts. Great game as always T90
This was an intense game lol