Gotta say I had never opened one of your videos being like “I hope my career as a particle theorist will help me understand any jokes he says”, but here we are.
As soon as I heard “Pokémon” and “worst speedrun” and “losing hours due to bad luck” and “attempted this too many times” I already knew it was Shen. What a brilliant mad lad.
Fun story: when I was like 9 my sister messed with my copy of Pokémon emerald, so I decided there was only one way to get her back. I took her copy of sapphire, leveled her combusken up to level 100 without letting it evolve, then left the game back in her room. It took so many hours since he was only level 26 and this was right before Wattson, but she was so upset she could never get her final evo Edit: spelling
Lol, a lvl 100 basic pokemon isnt necessarily "weaker" than its evolution. It just has a different set of stats, moves, and possibly types. Its essentially a different pokemon in the sense, not weaker or stronger. It all depends on how you utilize your pokemon.
@@lmercuryw the up and down quarks were named because they have a property called isospin, which behaves a lot like quantum mechanical spin in that there is an “up” and a “down” eigenvalue, thus the two particles associated with the up and down isospins were named the up and down quarks. The strange quark was named because it was literally strange; the hadrons that it composed (Kaons, mostly) did not have the expected value of isospin that most hadrons did, because they had one less up/down quark. Thus the new particle responsible for this “strangeness” quantity was named the strange quark. The charm quark was named as the antonym to strange, not too much thought into that. The top/bottom quarks were originally named truth and beauty, as a joke but it eventually just became the colloquial name for them in papers. They were both too large to be seen in any experiment at the time of positing, so it could have been that they thought the particles would be renamed once discovered. Top/bottom were chose eventually to just give them more normal names that had the same first letters (so earlier Feynman diagrams with little t’s and b’s wouldn’t be make obsolete)
One interesting wrinkle in the probability calculations at 8:15 is that they only check for the probability of getting *at least* one rare candy. Since what we actually care about is how long it takes to get enough Rare Candies (with the exception of not caring about overshooting, which only affects the odds at the end), the case where you get 2 Rare Candies simultaneously should be weighted twice as heavily. Accounting for this for each of 2-6 obtained, the average number of Rare Candies per fight in that that example... is exactly 0.06. The stated 5.85% number would closer to accurate if multiple Pickups couldn't proc on the same fight. If only one Pickup could proc per fight, the odds drop to 4.69% (as now picking up the wrong item aborts the attempt early)
I was wracking my brain of this. A Zigzagoon having a 1% chance to pick up a rare candy after the battle practically by definition means that over 100 battles that Zigzagoon will pick up 1 rare candy(on average), so naturally 6 Zigzagoons will pick up 6 rare candies over 100 battles. In other words, 6% of your battles will get you a rare candy.
There's kind of three layers to this: The first layer is the simple calculation - take the 1% per Zigzagoon and add it up - which is an incorrect method that would still yield the correct end result for what we care about. The second layer is what adef did, and calculating the probability of getting at least one rare candy. This isn't that hard to do - the odds of not getting a rare candy from a single Zigzagoon in a single battle is 99%, the odds of that happening 6 times is (0.99)^6 or 94.148...%, so the odds of 6 misses in a row not happening is one minus that, or the given 5.85%. The calculation is correct (adef did it by calculating the odds of getting 1, 2, 3, 4, 5, 6 and adding them up), but the thing being measured is incorrect. Which is the third layer - accounting for the fact that we care about how many rare candies are collected, not how many encounters award at least one rare candy. This results in an average of 0.06 candies per encounter, or as you put it, 6 candies per 100 encounters. Which is subtly different than 6% per encounter, even if it comes out to the correct value if you were to try to calculate an expected number of encounters to get enough rare candies to reach level 100. Probability is really, really hard.
in 5,8 battles you will pick up one candy, in 0,1 you would pick up 2, so together on average you pick up 0.06 per battle, but only in 0.058 you pick some up@@_Sebo
I wanted to see what the optimal strategy was for picking up items so, the goal is to optimize item rate. if a pokemon has a 10% chance to pick up an item, then each pokemon over n battles has a 1-(1-0.1)^n chance to have an item. Therefore, the expected number of items is 6(1-(1-0.1)^n). If we include the time taken to battle and the time taken to check the pokemon for items, we get the item rate = 6(1-(1-0.1)^n) / (time_battle * n + time_check) crudely timing from the video, with 16s for the battle and 4s to check items, the final equation is rate = 6(1-(1-0.1)^n) / (16 * n + 4) when we optimize for n, we get n=2.0982. So yeah, 2 is optimal.
I verified your result via simulation. Also, it is noteworthy that the result is independent of how many pokemon you have. It is probably better to check after every battle at the beginning, when battles are long... The breakpoint is somewhere around time_battle = 35s.
@@hhhhhh0175 This analysis included an assumption of 4seconds per item check, so in the analysis “Taking items off extra Pokémon” is assumed to be either negligible or included in the 4s of menuing. In my sim, increasing ALL menu time to 5s still results in n=2.
@@hhhhhh0175 No they don't as you're only checking IF they have an item, and then only proceeding about it IF there are ANY items. If you check and no one has an item, why are you wasting time in the summary screen?
@@SnoFitzroy and items aren't exactly easy to miss esp. on zig and lin: they're both brown and beige and the item's bright yellow and red. Literally a 1s glance, plus the menuing and fade in/out. (on that topic, I feel like someone should check the time for the menu in Ruby/Sapphire VS Emerald, to see if there's a load difference and an item take difference. I vaguely remember that sometimes one of them adds lag to menus but I'd have to go dig into a bunch of TASing threads to remember and my ADHD is already fighting me today I don't need a rabbit warren to wander off into)
Wouldn't it just be easier to calculate the odds of at least 1 rare candy pickup by multiplying 0.99 (the odds of no rare candy) to the power of 6, to get the odds of none of your six Zigzagoon getting a rare candy, and then subtracting that from 1.0? 0.99^6 = 0.94148, meaning a 5.852% of at least one rare candy pickup. Also yes, the charm joke was a good touch, as well as the idea of Pokemon ACT results or credit scores.
@@DrabekNewburn what do you mean "overly complex"? It's a lot simpler than summing each individual possibility of getting a given number of rare candies, that's for sure.
@@abtinbarzin8369 Pretty sure you're both in agreement and they're referring to the method shown in this video, which is a similar kind of summation method as used in the shiny hunting video. And I agree too (as much as I love these videos in general), not sure why the simpler method of subtracting the probability of failure from 1 isn't used in either video since it doesn't require any summations or probability distribution functions. adef even got several comments with feedback about this in the shiny-hunting video but it unfortunately seems like that feedback went ignored (or at least wasn't noticed).
Doing so technically doesn't account for the off scenario of multiple Rare Candy pickups happening at once (although granted, no answer providing a single probability really does). Which is highly unlikely, so ignoring that scenario is probably good enough for most purposes anyways.
Oh my god! Why would he do this?! What about the sanctity of this super arbitrary run performed over 4 years ago by a guy I've never met?! I care so much about this thing that I'm never gonna do and will probably never talk about with anyone ever again for the rest of my life!
The 6% simplification actually does work -- on average you get 0.06 rare candies per battle -- there is just a chance you get 2+ so the chance of 0 much be >6%. The party checking math feels definitely calculatable -- you would just need to know how long it takes check and how much time it saves by checking.
I suck at math but eyeballing it, wouldn't the ideal check be located somewhere between 5 to 8? I'd err on caution with going with 5 or pushing it with 6, but unless INCREDIBLY unlucky that shouldn't get you more than one or two proc'd pickups when you check, right?
@@neoqwerty Actually, two is pretty generically optimal unless your item checks are really slow (longer than your battle) or really fast (more than 9x faster than your battle).
what do you mean go with the vibe. there HAS to be a way to figure out an optimal frequency to check based on the expected amount of battles needed to spawn an item and a candy in particular, and the time loss it takes to check the party. there HAS TO *scribbles furiously*
Assuming every encounter gave max exp yield (34), it would take a minimum of 29412 battles to reach lv100. With 6 zigzagoon in the party, the probability would be given as (0.99^6)^29412 = 0.99^176472, a number so close to zero that if someone were to continuously do runs of this game from now until the heat death of the universe, it statistically would not happen even once. To put it in perspective, 0.99^229 is approximately 0.1 (10^-1), which means 0.99^458 is approximately 0.01 (10^-2), 0.99^687 is approximately 0.001 (10^-3), and so on. 0.99^176472 would be on the order of 10^-770. That's a decimal point followed by 770 zeroes before the first non-zero digit. Supposing you ran this game 1000 times per year, it would be 10^767 years before you're expected to see this happen even once. The heat death of the universe has been calculated as being on the order of 10^106 years. On average, it would take someone playing this game for 10^(767-106) = 10^661 heat deaths of the universe for this to happen even once.
hey, i just wanted to say thank you for making these videos. i’ve always told myself i was bad at math, but the way you explain it helps math make sense to me, and i really appreciate it!
I am one of the 10 or 11 people who laughed audibly when they heard Charm. Love this "series" as it was a desire to play the Pokemon TCG that jumpstarted my education. I had to learn to read and do basic math! Then it became calculating move potencies taking into account STAB and weakness/resistance...and now I'm having my students generate lines of best fit and run statistical analyses to see how much power creep there is or isn't in Pokemon :)
I did a "Level 100 before the first trainer" challenge as a kid in Ruby and Sapphire once, using 5 Zigzagoons plus my Torchic, grinding until I was, IIRC, level 33 with 67 rare candies so I could get to level 100. It took so ridiculously long that I got two shiny Poochyenas during the grind, I believe it was like 70 or so ingame hours. I thought this was such a terrible idea that only a kid with too much free time would do it. Never expected people to make a speedrun out of it.
In all honesty, I was super excited for you to go into the mathematically optimal amount of party checks for rare candies, I'm sure there's some really cool math behind that
Kind of feels like a missed opportunity to actually calculate the optimal number of encounters to trigger before checking your Pokemon's held items. Like, the data is there on pickup rate, and plenty of data could be acquired from the vods to figure out time expenditure on an encounter or on a menu check. How will we break the world record without knowing the optimal strategies? Great video as always.
Fun fact: The wild Pokemon's level is irrelevant in Gen 3 catching mechanics while using a normal Pokeball (Poke, great, and ultra; if the Zigzagoon were lv100 it would still be 78.5%
@10:31 Your video has got me thinking about the best way to mathematically decide how often one should check for rare candies! This run is too important to leave such a thing to chance ;)
11:09 Oh really? I thought we were going to use statistics for this one too. You can do that easily by calculating the diminishing return to find the optimal check rate based on the probabilities.
It's important to note for the 6% -> 5.85% distinction for "at least one rare candy" that what we're usually looking for is the expected value of rare candies, not the probability. And expected value _does_ behave additively, so every 100 battles with a full party you expect 6 rare candies (or "6% of a rare candy per battle" if you will). It also doesn't matter for time efficiency of when to check your inventory.
Hey, I just wanna say thanks for doing videos like this. I got a TBI from a car accident several years back and it kinda messed up my math abilities. I've really struggled with math since, and ever since then, I've tended to avoid it because it makes me really anxious. These videos make math somewhat more digestible for me, and hopefully will help me get back some of those skills I've lost without overwhelming me.
Long-run you’re expected to get 6 rare candies every 100 battles, which is functionally the same thing as 6% chance of getting 1. Great explainer on binomial distributions but for calculation/simulation purposes using a 6% hit rate works fine!
I'd really like someone calculating the optimal number of fights before checking your party, using the average time to check and average time for an encounter. And if they wanna go above and beyond, they can do this calc for every party size, as the numbers will be different. Kinda suprised shananagans apparently didn't do this himself. When you put yourself through the misery of this run, doing this optimization math seems like the fun part.
Dude you always cover the videos that I have researched in the past. I wish I had this video back then xD Trying to figure out why it wasn't 60% was time consuming for me. I knew it wasn't 60% but I coudn't figure out why.
i wrote up a little program to simulate this process with the goal of figuring out how often you should optimally check your party for items (among other things) but I need an estimate for how long it takes to complete a battle vs how long it takes to check your party and take out X number of items. (i think its fair to assume one tackle is enough to take out the enemy since this will be the case for most battles) BUT FOR NOW: Assuming battles and checking for items takes roughly the same amount of time (i have no idea but bear with me) we get the following values after a fairly large number of simulations ran until the level of the main zigzagoon + the number of rare candies equals 100: Checking after every battle: 1293 battles / 1293 item checks / 2586 total Checking every 2 battles: 1357 battles / 679 item checks / 2036 total Checking every 3 battles: 1410 battles / 471 item checks / 1881 total Checking every 4 battles: 1479 battles / 370 item checks / 1849 total Checking every 5 battles: 1539 battles / 308 item checks / 1847 total Which would suggest checking every 3/4/5 battles is roughly equal but all are significantly better than checking every second battle. Of course if in reality battles take longer than checking for items or the other way around this math changes, gonna need your guys' help with that part.
Only checking if there are any items can be done in just a couple seconds (Maybe around like 7 seconds? not sure about it), since you just need to open the bag and look for the held item icon on each pokemon. But if there are actually any held items when checking the party then having to menu to each one and remove the item would take probably around 5 seconds for every item.
@@th3bear01 perfect, now I just need to figure out how long a battle lasts, and I guess how long it takes to find a new battle turning around in the grass and then I can actually start getting time estimates
i might be wrong but the 5.85% chance is to get "at least 1 rare candy" on a given fight, but what its not accounting for is the chance includes getting multiple rare candies. the chance is miniscule to get 2 or 3, let alone any higher. If taken on the whole on a larger scale sample size however, the number of expected rare candies after 100 battles is indeed exactly 6. the "extra" 0.15% missing is "made up for" by the chance to get multiple rare candies from a single battle. But the small probabilities of getting multiple are "worth" twice or 3x as much to you, so you need to first multiply those percentages by 2 or 3, and when you do that for all 6 possibilities, you end up back at 6% total. This is a case where the simple answer is both incorrect and correct because the complicated answer makes it true.
The 5.7% is the chance of getting at least one candy. It's true however, that the expectation is still 0.06 . While probabilities cannot be added together, expectations can. After 100 battles, you will on average have 6 candies. The original comment is not wrong. Think about it like this. If you flip 2 coins, the expectation is that one of them will be heads. If you flip 100 coins, the expectation is that 50 of them are heads. The probability of precisely 50 heads will be very little as there's so many options between 40 and 60 that are also very likely to occur, but the expected value will still be 50, because expectations can be added together, and 50 is simply 100 times the expected value of a single coin flip, 0.5, just like in our 2 coin example 1 is 2 times the same expected value of 0.5 .
I love this content!! I was thinking about your channel the other day because Clay’s Excadrill in black 2 rock slide flinched me 5/6 times in 6 moves, including 4 in a row. According to my math being flinched 4 times in a row only has a 2.4% chance of happening without taking into account the move’s chance to miss. Crazy!!
Wait I think I did the math wrong. Here’s the updated work: Rock Slide has a 90% chance to hit, so hitting 6 in a row is .9^6 = .478 or a 47.8% chance of happening. The flinch chance is 30, which happened on 5/6 rock slides. So you’d do .3^4 (for the 4 flinches in a row) which is .0024 or (2.4% of happening) multiply it by .7 (for 70% chance of no flinch) to get .0017, then multiply that by .3 for the final flinch in 6 turns to get .00051 or .05% chance for 5/6 rock slide flinches. Then you’d take the .478 chance for the rock slide to hit 6 times in a row and multiply it by .00051 chance of 5/6 flinches and get .00024 or a .024% chance of that sequence happening. if anyone cares enough to read through that feel free to correct my math, but I felt like I was going crazy!!
@@MonJcFarland5that was honestly so good to read. Math isn’t my forte so nothing to correct. But, as a competitive player I really appreciate the odds, chances and mathematics that go into moves.
I did a similar thing on Ruby I think, but it was „Full Team Level 100 before 1st Gym“. Caught 6 Zigzagoons and killed the whole wildlife around with them. Then Rare Candied Nincada to 99, and let it evolve on 100 to safe a bit of grinding before. Took a whole lot of time, but also was fun to have an unbeatable team straight from the start.
I had always heard that you can do additive probabilities by determining the chance of the desired outcome not happening, then inverting. The chance of a thing happening at least once over X tries is the same as one minus the chance of it not happening X times consecutively. So the odds of one or more rare candy pickups for six ziggies is 1-(0.99^6)=5.852%, if that method is valid.
Fun detail about the math part of determining the odds of getting a rare candy, it's much easier to determine the odds of not getting any rare candies (binomial equation with x=0 which effectively renders the whole equation to P = .99^6 = 0.94148...) and then subtract that from 1, you get the same result which is just the odds of not getting none! I got very tired of running multiple layers of the equation in high school every time i got a "probability of at least one of X" question, and pretty quickly figured out that "odds of at least one" is the same as "odds of not none" which is much faster and easier to calculate :D
Now we wait for someone to be crazy enough to do this in an unmodified game with just their starter (the heat death of the universe will come before they finish the "speedrun")
There's definitely some math to be done with how often you should check the party. I think it ultimately boils down to average pickup rolls per unit of time, because the expected value of rare candies you get is directly proportional to the pickup rolls you have performed. The expected value for pickup activating with 6 Zigzagoons is 0.6 per battle. If the team isn't checked in between, in the next battle there will only be an expected number of 5.4 rolls, so the expected value of activations is 0.54. After that, these numbers are 4.86 and 0.486, then 4.374 and 0.4374, and so on. The effectiveness of each battle reduces by 10% compared to the previous one if the team isn't checked in between. Now, someone of course would need to determine how big proportion of time checking the team takes on average in a cycle of *run in grass -> battle -> check team -> repeat*. With this knowledge, an optimal rate of checks could be determined based on how much time is saved by skipping checks vs how much pickup effectiveness is lost by not checking the team. For what it's worth, my guess would be that checking the team every two battles is probably not too far from optimal, or it even might be optimal. Because I don't think checking it takes a big portion of time in the cycle.
You could also calculate how much time was saved by the chat speedup too, by taking the time difference between normal and the speed per battle * the number of battles in a run. If someone were willing to count how many battles there were that is.
I was wondering how RNG could vary full hours until you said the word "Rare Candy" and it all made sense. Fantastic video. As a math nerd and Pokemon nerd, I'm really glad I'm finally checking you out! EDIT: Oh god and the Monkey Ball music.
I mean for the text thing he’s already a modified game to do the speedrun. So long as all other runners have access to the same game modified the same way it’s completely fair
I was kinda expecting u to tell us what the optimal number of battles between checking items was, with your probability magic though :( You'd need to list the time it takes to check inv vs the time it takes to do 1 battle (on average) and compare the time you lose on checking vs the time you'd lose from probability bleed of already having 1 or more items (you could totally say you're losing 1/6 of a battle's time if your zizagons are holding 1 item and so on + (1/6)/x of the checking time, being x the number of battles you do before checking) It'd be a fun spreadsheet to work! ;D
This reminds me of the speedrun where a guy grinded (ground?) Cloud and Barrett to level cap in the first area of FF7. Took literally hundreds of hours, as you might expect.
I remember Shen's video. I have the VOD on my backlog after I discovered it two or so years back. I'm about halfway through it after passively listening to it on and off.
The funny thing is that the total number of rare candies generated on average can be calculated just fine by ignoring the binomial theorem stuff and just saying 6% total :P
Wow, I used a VOD of one of these runs to fall asleep almost every night for months a few years ago. Wouldn't be surprised if like 10% of the views on the video were just mine haha. It was just so soothing to watch and I'm really grateful for Shen for putting himself through grinding hell for our sake.
Although the probability of getting a rare candy is indeed less than 6%, the average number of rare candies per battle is exactly 6%. Assuming you remove items after every battle, you would find an average of 6 rare candies per 100 battles. Edit: Also, there is probably a mathematical answer to how battles you should do before checking for items, but it would depend on how long it takes to check vs how much time is wasted if you don't.
Thanks for another great video! Is having Medium Fast there twice also part of of the quantum joke? Also, I'm a bit sad you didn't go further into the maths of when to check - or at least how much the odds change if one or two pick up an item. I'd love to know at least.
I expected a little bit more optimization theory about how often to check your party, so here's my back of the napkin calculations. The expected efficiency loss you take by not checking your party is the difference in expected value of battle 2 compared to battle 1, based off of the 10% item pickup rate. That EV comes out to just be the number of zigzagoons times 10%, so .6 with full party, .5 if one zigzagoon has an item, etc. After the first battle, there's a 35% chance that one zigzagoon now has a full inventory, 10% chance two do, and a small tail that I'll ignore. So 35% of the time, the EV for battle 2 drops to .5, and 10% of the time it drops to .4, meaning the EV is .5*.35+.4*.1+.6*.55=5.45. I think you can actually just say EV(n)=.9EV(n-1), and thus EV(n)=.9^n*EV(0). If that's correct, which i'm too lazy to check, the optimization problem is just checking when those two are equal. Say that the time it takes to do a battle is 1, while the time it takes to check your party is p, and the battles you run before checking party is b. Your time-loss inefficiency is p/b percent, while your EV loss inefficiency is EV(b)=(.9^b)*.6. Say it takes the same time to check your party as it does to fight one battle. p=1, so left side becomes 1/b, and right side is (.9^b)*.6. Those intersect at just over b=2, so checking every other battle does actually seem to be optimal.
I got the quark joke immediately (like. went "charm? like the quark?" before you even mentioned it) and I *do* have humor in my life, thank you very much. Yes that humor is me but it's something
I think there should be added to this de facto "Zigzagoon lvl 100" category the necessity for it to be a shiny Zigzagoon that reaches lvl 100, it would be funnier :3
Hot take: instant text speed is AMAZING. Every rpg with lots of text boxes should have an option for instant or really fast text speed and here's why: I can argue that in any rpg, depending on how much you like reading and how much crucial information you need to succeed, up to 60%-95% of all text in text boxes is completely useless to you, even in the first playthrough. It's all fluff. And that's fine, it gives characters personality, there's a plot to be told, perhaps some jokes to be made. But of all the info in text boxes and all the info the player needs to have fun or to progress, most of it is just useless. So most of the time when I'm playing a game with tons of text boxes I'll end up either skipping it or read it really fast and keep going
if nothing else all games should have an opt-in instant text (opt-in so the average "I'LL NEVER GO IN THE OPTIONS MENU EVERRRRR" gamer who also doesn't read the fricking manual can't then go yell at the devs when they inevitably get lost because they skipped over the plot that went "yo go fetch this item from this mysterious desert dungeon") AND/OR have it on by default in NewGame+.
Thats the best Brilliant ad I've seen. Aside from that, damn I still don't understand how you manage to make the most boring subject ever so frickin interesting. Seeing a new ADEF video in my notifs is always a rush 😂
"Unless you want to, I don't know, I'm not your dad." In before years from now when adef's future kid stumbles upon this video and decides to attempt the run. Also medium fast was listed on the xp curves list twice.
So I've been thinking that you sound like someone else I listen to, and I couldn't quite put my finger on it until just now. Your voice and inflections sound SO MUCH like stand-up comedian Drew Lynch, but without the stutter. Also, the Charm joke was cute. I slapped my knee, even!
I ... I just watched a Brilliant ad at the end. The whole thing. adef is completely fired up, take care guys and watch out for an unhinged mathematician i guess. i´m not that deep into adef lore
Since we care about the number of rare candies received rather than whether pickup activated, the expected value is more relevant. And you can always just add expected values together, even for events that aren't independent.
ziggy
also insta text is cheating
You and werster are real pokemon legends.
Nothing but respect for the sheer amount of patience you have
I'm only here for the subtle nods towards quantum physics.
What stats did you end up with? What was average battle time or time between battles?
Gotta say I had never opened one of your videos being like “I hope my career as a particle theorist will help me understand any jokes he says”, but here we are.
It's been a while since I had a physics class - was that a quark joke? I remember charming/weird being a dichotomy with those
@@Omnicide101 yes, the charm quark is one of the 6! Its mass was also put on screen instead of experience.
I read particle terrorist :0
Me, an English major who watches BobbyBroccoli:
I thought that was funny too when he had the giga electronvolts / c^2
As soon as I heard “Pokémon” and “worst speedrun” and “losing hours due to bad luck” and “attempted this too many times” I already knew it was Shen. What a brilliant mad lad.
Fun story: when I was like 9 my sister messed with my copy of Pokémon emerald, so I decided there was only one way to get her back. I took her copy of sapphire, leveled her combusken up to level 100 without letting it evolve, then left the game back in her room. It took so many hours since he was only level 26 and this was right before Wattson, but she was so upset she could never get her final evo
Edit: spelling
That's genuinely evil
Lol, a lvl 100 basic pokemon isnt necessarily "weaker" than its evolution. It just has a different set of stats, moves, and possibly types. Its essentially a different pokemon in the sense, not weaker or stronger. It all depends on how you utilize your pokemon.
Honestly, game shark was such a good thing, man,
@@fantasystaplesuwu1554 Uh... no, a lv100 Combusken is strictly weaker than a lv100 Blaziken.
@@fantasystaplesuwu1554 "All of the base stats are lower, but it's not weaker."
Lol. Lmao even.
Here's me thinking this was a one time lol meme run Shen did and then you said "his previous time". This maniac has done this more than once?!?
I have never in my life been more disappointed at not getting a joke than I am with Charm.
Same. Now I have to pivot to become a math major...
It's one of the components of an atom, to vastly simplify.
it’s a quark. quarks are in pairs, which are: top/bottom, up/down, and charm/strange. dont know what those physicists were smoking lol
And it's mass is listed in natural units in the total experience table
@@lmercuryw the up and down quarks were named because they have a property called isospin, which behaves a lot like quantum mechanical spin in that there is an “up” and a “down” eigenvalue, thus the two particles associated with the up and down isospins were named the up and down quarks.
The strange quark was named because it was literally strange; the hadrons that it composed (Kaons, mostly) did not have the expected value of isospin that most hadrons did, because they had one less up/down quark. Thus the new particle responsible for this “strangeness” quantity was named the strange quark.
The charm quark was named as the antonym to strange, not too much thought into that.
The top/bottom quarks were originally named truth and beauty, as a joke but it eventually just became the colloquial name for them in papers. They were both too large to be seen in any experiment at the time of positing, so it could have been that they thought the particles would be renamed once discovered. Top/bottom were chose eventually to just give them more normal names that had the same first letters (so earlier Feynman diagrams with little t’s and b’s wouldn’t be make obsolete)
lol he's so 'quarky'
If it werent for the fact that Adef’s younger than me, i’d think he WAS my dad
How old is he??
@@fosterdawson7339 He's 32
34
I'm your dad
Hell I thought he was me until I realized he isn’t
That charm joke was a bit... strange... not over the Top but not rock Bottom either. We all have our Ups and Downs 🤓
I’m sure this is really funny so I’m just trusting it makes sense and giving it a like
@@discosolo These are the names of the types of quarks! Charm, strange, top, bottom, up, and down.
ya beat me too it, gg
I made a charm joke, but this was too good. I deleted it. WELL DONE!
It's an acquired flavor
One interesting wrinkle in the probability calculations at 8:15 is that they only check for the probability of getting *at least* one rare candy. Since what we actually care about is how long it takes to get enough Rare Candies (with the exception of not caring about overshooting, which only affects the odds at the end), the case where you get 2 Rare Candies simultaneously should be weighted twice as heavily. Accounting for this for each of 2-6 obtained, the average number of Rare Candies per fight in that that example... is exactly 0.06.
The stated 5.85% number would closer to accurate if multiple Pickups couldn't proc on the same fight. If only one Pickup could proc per fight, the odds drop to 4.69% (as now picking up the wrong item aborts the attempt early)
I was wracking my brain of this. A Zigzagoon having a 1% chance to pick up a rare candy after the battle practically by definition means that over 100 battles that Zigzagoon will pick up 1 rare candy(on average), so naturally 6 Zigzagoons will pick up 6 rare candies over 100 battles. In other words, 6% of your battles will get you a rare candy.
There's kind of three layers to this:
The first layer is the simple calculation - take the 1% per Zigzagoon and add it up - which is an incorrect method that would still yield the correct end result for what we care about.
The second layer is what adef did, and calculating the probability of getting at least one rare candy. This isn't that hard to do - the odds of not getting a rare candy from a single Zigzagoon in a single battle is 99%, the odds of that happening 6 times is (0.99)^6 or 94.148...%, so the odds of 6 misses in a row not happening is one minus that, or the given 5.85%. The calculation is correct (adef did it by calculating the odds of getting 1, 2, 3, 4, 5, 6 and adding them up), but the thing being measured is incorrect.
Which is the third layer - accounting for the fact that we care about how many rare candies are collected, not how many encounters award at least one rare candy. This results in an average of 0.06 candies per encounter, or as you put it, 6 candies per 100 encounters. Which is subtly different than 6% per encounter, even if it comes out to the correct value if you were to try to calculate an expected number of encounters to get enough rare candies to reach level 100.
Probability is really, really hard.
in 5,8 battles you will pick up one candy, in 0,1 you would pick up 2, so together on average you pick up 0.06 per battle, but only in 0.058 you pick some up@@_Sebo
The "let them have the charm joke, they need it" got me
I wanted to see what the optimal strategy was for picking up items
so, the goal is to optimize item rate.
if a pokemon has a 10% chance to pick up an item, then each pokemon over n battles has a 1-(1-0.1)^n chance to have an item.
Therefore, the expected number of items is 6(1-(1-0.1)^n).
If we include the time taken to battle and the time taken to check the pokemon for items, we get the item rate =
6(1-(1-0.1)^n) / (time_battle * n + time_check)
crudely timing from the video, with 16s for the battle and 4s to check items, the final equation is
rate = 6(1-(1-0.1)^n) / (16 * n + 4)
when we optimize for n, we get n=2.0982.
So yeah, 2 is optimal.
I verified your result via simulation.
Also, it is noteworthy that the result is independent of how many pokemon you have.
It is probably better to check after every battle at the beginning, when battles are long... The breakpoint is somewhere around time_battle = 35s.
it's not independent, more pokemon take longer to check in the menu
@@hhhhhh0175 This analysis included an assumption of 4seconds per item check, so in the analysis “Taking items off extra Pokémon” is assumed to be either negligible or included in the 4s of menuing.
In my sim, increasing ALL menu time to 5s still results in n=2.
@@hhhhhh0175 No they don't as you're only checking IF they have an item, and then only proceeding about it IF there are ANY items. If you check and no one has an item, why are you wasting time in the summary screen?
@@SnoFitzroy and items aren't exactly easy to miss esp. on zig and lin: they're both brown and beige and the item's bright yellow and red. Literally a 1s glance, plus the menuing and fade in/out.
(on that topic, I feel like someone should check the time for the menu in Ruby/Sapphire VS Emerald, to see if there's a load difference and an item take difference. I vaguely remember that sometimes one of them adds lag to menus but I'd have to go dig into a bunch of TASing threads to remember and my ADHD is already fighting me today I don't need a rabbit warren to wander off into)
Wouldn't it just be easier to calculate the odds of at least 1 rare candy pickup by multiplying 0.99 (the odds of no rare candy) to the power of 6, to get the odds of none of your six Zigzagoon getting a rare candy, and then subtracting that from 1.0? 0.99^6 = 0.94148, meaning a 5.852% of at least one rare candy pickup.
Also yes, the charm joke was a good touch, as well as the idea of Pokemon ACT results or credit scores.
Same overly-complex "solution" as in the last video.
@@DrabekNewburn what do you mean "overly complex"? It's a lot simpler than summing each individual possibility of getting a given number of rare candies, that's for sure.
yes, but the binomial coefficients are exactly that symmetrical, so that will come out to the same.
@@abtinbarzin8369 Pretty sure you're both in agreement and they're referring to the method shown in this video, which is a similar kind of summation method as used in the shiny hunting video.
And I agree too (as much as I love these videos in general), not sure why the simpler method of subtracting the probability of failure from 1 isn't used in either video since it doesn't require any summations or probability distribution functions. adef even got several comments with feedback about this in the shiny-hunting video but it unfortunately seems like that feedback went ignored (or at least wasn't noticed).
Doing so technically doesn't account for the off scenario of multiple Rare Candy pickups happening at once (although granted, no answer providing a single probability really does).
Which is highly unlikely, so ignoring that scenario is probably good enough for most purposes anyways.
Oh my god! Why would he do this?! What about the sanctity of this super arbitrary run performed over 4 years ago by a guy I've never met?! I care so much about this thing that I'm never gonna do and will probably never talk about with anyone ever again for the rest of my life!
Oh wait, oh it doesn't matter
🎶 but in the end, it doesn't even matter 🎶
We found a new pasta to paste when someone does this lmao!
@@randomgamer-te8op♪ I had to fall, to lose it all ♪
@@Artoosa i mean this applies to basically every youtube video ever
1:55 It's not cheating, it's microtransactions in 2006 baby.
The 6% simplification actually does work -- on average you get 0.06 rare candies per battle -- there is just a chance you get 2+ so the chance of 0 much be >6%.
The party checking math feels definitely calculatable -- you would just need to know how long it takes check and how much time it saves by checking.
I suck at math but eyeballing it, wouldn't the ideal check be located somewhere between 5 to 8? I'd err on caution with going with 5 or pushing it with 6, but unless INCREDIBLY unlucky that shouldn't get you more than one or two proc'd pickups when you check, right?
@@neoqwerty Actually, two is pretty generically optimal unless your item checks are really slow (longer than your battle) or really fast (more than 9x faster than your battle).
what do you mean go with the vibe. there HAS to be a way to figure out an optimal frequency to check based on the expected amount of battles needed to spawn an item and a candy in particular, and the time loss it takes to check the party. there HAS TO *scribbles furiously*
The charm joke was indeed very funny
Still need to find the other 10 guys
I hate that I got it
@@Laezar1I hate that I didn't get it immediately.
It took only a few seconds, granted.
Oh please none of yall got it 😂
@@atomdecay its a quark joke a 2nd yr undergrad would get it
Its the pokemon math twink again, yay!
Engagement, baby!
Not the pokemon math twink 😂
"big brain = good job"
*cries in working retail with a software design degree*
Software design is a fake degree
Got my Environmental Science Masters 2 years ago and have yet to find a job in the field. I feel you pal.
Laughs in no degree, works in software
The real question is how unlikely is it for him to reach level 100 without getting any rare candies the entire time.
Assuming every encounter gave max exp yield (34), it would take a minimum of 29412 battles to reach lv100. With 6 zigzagoon in the party, the probability would be given as (0.99^6)^29412 = 0.99^176472, a number so close to zero that if someone were to continuously do runs of this game from now until the heat death of the universe, it statistically would not happen even once.
To put it in perspective, 0.99^229 is approximately 0.1 (10^-1), which means 0.99^458 is approximately 0.01 (10^-2), 0.99^687 is approximately 0.001 (10^-3), and so on. 0.99^176472 would be on the order of 10^-770. That's a decimal point followed by 770 zeroes before the first non-zero digit. Supposing you ran this game 1000 times per year, it would be 10^767 years before you're expected to see this happen even once.
The heat death of the universe has been calculated as being on the order of 10^106 years. On average, it would take someone playing this game for 10^(767-106) = 10^661 heat deaths of the universe for this to happen even once.
hey, i just wanted to say thank you for making these videos. i’ve always told myself i was bad at math, but the way you explain it helps math make sense to me, and i really appreciate it!
I am one of the 10 or 11 people who laughed audibly when they heard Charm.
Love this "series" as it was a desire to play the Pokemon TCG that jumpstarted my education. I had to learn to read and do basic math! Then it became calculating move potencies taking into account STAB and weakness/resistance...and now I'm having my students generate lines of best fit and run statistical analyses to see how much power creep there is or isn't in Pokemon :)
I did a "Level 100 before the first trainer" challenge as a kid in Ruby and Sapphire once, using 5 Zigzagoons plus my Torchic, grinding until I was, IIRC, level 33 with 67 rare candies so I could get to level 100. It took so ridiculously long that I got two shiny Poochyenas during the grind, I believe it was like 70 or so ingame hours. I thought this was such a terrible idea that only a kid with too much free time would do it. Never expected people to make a speedrun out of it.
how did u get the zigs? :0
In all honesty, I was super excited for you to go into the mathematically optimal amount of party checks for rare candies, I'm sure there's some really cool math behind that
i like this channel because even though i dont like math, your choices of topic make me remember how much i enjoyed math until my final year in school
Kind of feels like a missed opportunity to actually calculate the optimal number of encounters to trigger before checking your Pokemon's held items. Like, the data is there on pickup rate, and plenty of data could be acquired from the vods to figure out time expenditure on an encounter or on a menu check.
How will we break the world record without knowing the optimal strategies?
Great video as always.
another day, another adef video where I keep pausing and stopping + starting the video to catch second long text jokes
3:54 lol, it's its mass-energy equivalence!
Fun fact: The wild Pokemon's level is irrelevant in Gen 3 catching mechanics while using a normal Pokeball (Poke, great, and ultra; if the Zigzagoon were lv100 it would still be 78.5%
You explained the Binomial Distribution better than my college professors did
Math teachers should learn to choose a topic we really care about!
@10:31 Your video has got me thinking about the best way to mathematically decide how often one should check for rare candies! This run is too important to leave such a thing to chance ;)
11:09 Oh really? I thought we were going to use statistics for this one too. You can do that easily by calculating the diminishing return to find the optimal check rate based on the probabilities.
Why don't you do it then
It's important to note for the 6% -> 5.85% distinction for "at least one rare candy" that what we're usually looking for is the expected value of rare candies, not the probability. And expected value _does_ behave additively, so every 100 battles with a full party you expect 6 rare candies (or "6% of a rare candy per battle" if you will). It also doesn't matter for time efficiency of when to check your inventory.
Hey, I just wanna say thanks for doing videos like this. I got a TBI from a car accident several years back and it kinda messed up my math abilities. I've really struggled with math since, and ever since then, I've tended to avoid it because it makes me really anxious. These videos make math somewhat more digestible for me, and hopefully will help me get back some of those skills I've lost without overwhelming me.
I love that Zigzagoon is getting more love and attention because of a super strange speed run cause Zigzagoon is awesome
Long-run you’re expected to get 6 rare candies every 100 battles, which is functionally the same thing as 6% chance of getting 1.
Great explainer on binomial distributions but for calculation/simulation purposes using a 6% hit rate works fine!
binomial distribution is great if you just care that an event happens, but here we care how many candys we get, so expected value is actually better.
I'm SO curious how a TAS of this would go.
I'd really like someone calculating the optimal number of fights before checking your party, using the average time to check and average time for an encounter. And if they wanna go above and beyond, they can do this calc for every party size, as the numbers will be different.
Kinda suprised shananagans apparently didn't do this himself. When you put yourself through the misery of this run, doing this optimization math seems like the fun part.
Dude you always cover the videos that I have researched in the past. I wish I had this video back then xD Trying to figure out why it wasn't 60% was time consuming for me. I knew it wasn't 60% but I coudn't figure out why.
I cant wait to sort my pokemon by their, Birth Sign, Rising Sign, Birth Stone, ACT Score, PSAT Score, Myers-Briggs and their Credit Rating!
i wrote up a little program to simulate this process with the goal of figuring out how often you should optimally check your party for items (among other things) but I need an estimate for how long it takes to complete a battle vs how long it takes to check your party and take out X number of items.
(i think its fair to assume one tackle is enough to take out the enemy since this will be the case for most battles)
BUT FOR NOW:
Assuming battles and checking for items takes roughly the same amount of time (i have no idea but bear with me) we get the following values after a fairly large number of simulations ran until the level of the main zigzagoon + the number of rare candies equals 100:
Checking after every battle: 1293 battles / 1293 item checks / 2586 total
Checking every 2 battles: 1357 battles / 679 item checks / 2036 total
Checking every 3 battles: 1410 battles / 471 item checks / 1881 total
Checking every 4 battles: 1479 battles / 370 item checks / 1849 total
Checking every 5 battles: 1539 battles / 308 item checks / 1847 total
Which would suggest checking every 3/4/5 battles is roughly equal but all are significantly better than checking every second battle. Of course if in reality battles take longer than checking for items or the other way around this math changes, gonna need your guys' help with that part.
Only checking if there are any items can be done in just a couple seconds (Maybe around like 7 seconds? not sure about it), since you just need to open the bag and look for the held item icon on each pokemon. But if there are actually any held items when checking the party then having to menu to each one and remove the item would take probably around 5 seconds for every item.
@@th3bear01 perfect, now I just need to figure out how long a battle lasts, and I guess how long it takes to find a new battle turning around in the grass and then I can actually start getting time estimates
Shen's crazy meme categories are my favorite thing.
i might be wrong but the 5.85% chance is to get "at least 1 rare candy" on a given fight, but what its not accounting for is the chance includes getting multiple rare candies. the chance is miniscule to get 2 or 3, let alone any higher. If taken on the whole on a larger scale sample size however, the number of expected rare candies after 100 battles is indeed exactly 6. the "extra" 0.15% missing is "made up for" by the chance to get multiple rare candies from a single battle. But the small probabilities of getting multiple are "worth" twice or 3x as much to you, so you need to first multiply those percentages by 2 or 3, and when you do that for all 6 possibilities, you end up back at 6% total. This is a case where the simple answer is both incorrect and correct because the complicated answer makes it true.
No, the binomial distribution gets the correct result, like he said. Your assumption of 1-(1-.01)^6 actually gives 5.7%, like he said in the video.
The 5.7% is the chance of getting at least one candy. It's true however, that the expectation is still 0.06 . While probabilities cannot be added together, expectations can. After 100 battles, you will on average have 6 candies. The original comment is not wrong. Think about it like this. If you flip 2 coins, the expectation is that one of them will be heads. If you flip 100 coins, the expectation is that 50 of them are heads. The probability of precisely 50 heads will be very little as there's so many options between 40 and 60 that are also very likely to occur, but the expected value will still be 50, because expectations can be added together, and 50 is simply 100 times the expected value of a single coin flip, 0.5, just like in our 2 coin example 1 is 2 times the same expected value of 0.5 .
@@genessab ua-cam.com/video/hXEY0bG0Yew/v-deo.html This video has a good explainer of the same concept
I love this content!! I was thinking about your channel the other day because Clay’s Excadrill in black 2 rock slide flinched me 5/6 times in 6 moves, including 4 in a row. According to my math being flinched 4 times in a row only has a 2.4% chance of happening without taking into account the move’s chance to miss. Crazy!!
Wait I think I did the math wrong. Here’s the updated work: Rock Slide has a 90% chance to hit, so hitting 6 in a row is .9^6 = .478 or a 47.8% chance of happening. The flinch chance is 30, which happened on 5/6 rock slides. So you’d do .3^4 (for the 4 flinches in a row) which is .0024 or (2.4% of happening) multiply it by .7 (for 70% chance of no flinch) to get .0017, then multiply that by .3 for the final flinch in 6 turns to get .00051 or .05% chance for 5/6 rock slide flinches. Then you’d take the .478 chance for the rock slide to hit 6 times in a row and multiply it by .00051 chance of 5/6 flinches and get .00024 or a .024% chance of that sequence happening. if anyone cares enough to read through that feel free to correct my math, but I felt like I was going crazy!!
@@MonJcFarland5that was honestly so good to read. Math isn’t my forte so nothing to correct. But, as a competitive player I really appreciate the odds, chances and mathematics that go into moves.
Crazy, what an electric bogaloo. Speedruners show an amaizing dedication, and adef sure had a lot of dedication with this video
I did a similar thing on Ruby I think, but it was „Full Team Level 100 before 1st Gym“. Caught 6 Zigzagoons and killed the whole wildlife around with them. Then Rare Candied Nincada to 99, and let it evolve on 100 to safe a bit of grinding before. Took a whole lot of time, but also was fun to have an unbeatable team straight from the start.
3:45 it's a very Strange joke, at that
13:35 nice
I had always heard that you can do additive probabilities by determining the chance of the desired outcome not happening, then inverting. The chance of a thing happening at least once over X tries is the same as one minus the chance of it not happening X times consecutively. So the odds of one or more rare candy pickups for six ziggies is 1-(0.99^6)=5.852%, if that method is valid.
Nice video. You touched on it a bit, but I would love to see a deeper dive into the optimal frequency at which to check your party for items.
Fun detail about the math part of determining the odds of getting a rare candy, it's much easier to determine the odds of not getting any rare candies (binomial equation with x=0 which effectively renders the whole equation to P = .99^6 = 0.94148...) and then subtract that from 1, you get the same result which is just the odds of not getting none!
I got very tired of running multiple layers of the equation in high school every time i got a "probability of at least one of X" question, and pretty quickly figured out that "odds of at least one" is the same as "odds of not none" which is much faster and easier to calculate :D
3:34 Why was medium fast on the list twice?
Medium Fast… …more like Twin Medium Fast.
To calc the probability of not finding a rare candy with 6 zigzagoons in the party you can also just 1-0.99^6
Now we wait for someone to be crazy enough to do this in an unmodified game with just their starter (the heat death of the universe will come before they finish the "speedrun")
Don't make me
Cause I will
I spent 3 years getting a No Damage 100% run in the hardest Castlevania, I don't run much but don't make me do this cause I will
4 minutes in I realized realized zigzagoon has pickup which can give rare candies. Oh no. This would be torture.
8:18 1-((1-p)^6) for p=0.01 also works. The probability of not all failures.
This is how I thought about it, and I think it's more intuitive imo
There's definitely some math to be done with how often you should check the party. I think it ultimately boils down to average pickup rolls per unit of time, because the expected value of rare candies you get is directly proportional to the pickup rolls you have performed.
The expected value for pickup activating with 6 Zigzagoons is 0.6 per battle. If the team isn't checked in between, in the next battle there will only be an expected number of 5.4 rolls, so the expected value of activations is 0.54. After that, these numbers are 4.86 and 0.486, then 4.374 and 0.4374, and so on. The effectiveness of each battle reduces by 10% compared to the previous one if the team isn't checked in between.
Now, someone of course would need to determine how big proportion of time checking the team takes on average in a cycle of *run in grass -> battle -> check team -> repeat*. With this knowledge, an optimal rate of checks could be determined based on how much time is saved by skipping checks vs how much pickup effectiveness is lost by not checking the team. For what it's worth, my guess would be that checking the team every two battles is probably not too far from optimal, or it even might be optimal. Because I don't think checking it takes a big portion of time in the cycle.
The moment you mention Rare Candies I felt immediate dread upon the thought of pickup grind
You could also calculate how much time was saved by the chat speedup too, by taking the time difference between normal and the speed per battle * the number of battles in a run. If someone were willing to count how many battles there were that is.
I thought the charm joke was clairo shade for a moment. Another banger, thanks adef
I was wondering how RNG could vary full hours until you said the word "Rare Candy" and it all made sense.
Fantastic video. As a math nerd and Pokemon nerd, I'm really glad I'm finally checking you out!
EDIT: Oh god and the Monkey Ball music.
I got the Charm joke and it was very much appreciated.
Fantastic video!!
I mean for the text thing he’s already a modified game to do the speedrun. So long as all other runners have access to the same game modified the same way it’s completely fair
You know what would be really cool. If you could calculate how much time it takes to get theese 69 candies basing on how long a fight takes.
I was kinda expecting u to tell us what the optimal number of battles between checking items was, with your probability magic though :(
You'd need to list the time it takes to check inv vs the time it takes to do 1 battle (on average) and compare the time you lose on checking vs the time you'd lose from probability bleed of already having 1 or more items (you could totally say you're losing 1/6 of a battle's time if your zizagons are holding 1 item and so on + (1/6)/x of the checking time, being x the number of battles you do before checking)
It'd be a fun spreadsheet to work! ;D
This reminds me of the speedrun where a guy grinded (ground?) Cloud and Barrett to level cap in the first area of FF7. Took literally hundreds of hours, as you might expect.
Listing the charm mass in the table was a great touch
Something horrible that crossed my mind was doing this in GSC. No pick up ability would probably take 20+ hrs
I've actually seen a run like this, not as bad as you'd expect. You catch raikou and then get 60 rare candies from buena's pasword
2020… Why does that year sound vaguely familiar for some reason?
This year is cursed asf
I remember Shen's video. I have the VOD on my backlog after I discovered it two or so years back. I'm about halfway through it after passively listening to it on and off.
Note: in the experience groups table, you listed 'medium fast' twice.
The funny thing is that the total number of rare candies generated on average can be calculated just fine by ignoring the binomial theorem stuff and just saying 6% total :P
Bro turned his Zigzagoon party into a factory
Wow, I used a VOD of one of these runs to fall asleep almost every night for months a few years ago. Wouldn't be surprised if like 10% of the views on the video were just mine haha. It was just so soothing to watch and I'm really grateful for Shen for putting himself through grinding hell for our sake.
Although the probability of getting a rare candy is indeed less than 6%, the average number of rare candies per battle is exactly 6%.
Assuming you remove items after every battle, you would find an average of 6 rare candies per 100 battles.
Edit: Also, there is probably a mathematical answer to how battles you should do before checking for items, but it would depend on how long it takes to check vs how much time is wasted if you don't.
Thanks for another great video! Is having Medium Fast there twice also part of of the quantum joke?
Also, I'm a bit sad you didn't go further into the maths of when to check - or at least how much the odds change if one or two pick up an item. I'd love to know at least.
Real quarky joke you put in there
charm joke awakened my high school physics knowledge from a deep sleep
I expected a little bit more optimization theory about how often to check your party, so here's my back of the napkin calculations.
The expected efficiency loss you take by not checking your party is the difference in expected value of battle 2 compared to battle 1, based off of the 10% item pickup rate. That EV comes out to just be the number of zigzagoons times 10%, so .6 with full party, .5 if one zigzagoon has an item, etc.
After the first battle, there's a 35% chance that one zigzagoon now has a full inventory, 10% chance two do, and a small tail that I'll ignore. So 35% of the time, the EV for battle 2 drops to .5, and 10% of the time it drops to .4, meaning the EV is .5*.35+.4*.1+.6*.55=5.45. I think you can actually just say EV(n)=.9EV(n-1), and thus EV(n)=.9^n*EV(0).
If that's correct, which i'm too lazy to check, the optimization problem is just checking when those two are equal. Say that the time it takes to do a battle is 1, while the time it takes to check your party is p, and the battles you run before checking party is b. Your time-loss inefficiency is p/b percent, while your EV loss inefficiency is EV(b)=(.9^b)*.6.
Say it takes the same time to check your party as it does to fight one battle.
p=1, so left side becomes 1/b, and right side is (.9^b)*.6. Those intersect at just over b=2, so checking every other battle does actually seem to be optimal.
I love using Earthbound fight backgrounds to keep the screen visually interesting
The Charm bit got me.
Not a physicist, just a Hank Green enthusiast.
I got the quark joke immediately (like. went "charm? like the quark?" before you even mentioned it) and I *do* have humor in my life, thank you very much.
Yes that humor is me but it's something
I think there should be added to this de facto "Zigzagoon lvl 100" category the necessity for it to be a shiny Zigzagoon that reaches lvl 100, it would be funnier :3
That was the first shen stream I ever watched. Crazy moment in my life for no other reason than that
Hot take: instant text speed is AMAZING. Every rpg with lots of text boxes should have an option for instant or really fast text speed and here's why:
I can argue that in any rpg, depending on how much you like reading and how much crucial information you need to succeed, up to 60%-95% of all text in text boxes is completely useless to you, even in the first playthrough. It's all fluff. And that's fine, it gives characters personality, there's a plot to be told, perhaps some jokes to be made. But of all the info in text boxes and all the info the player needs to have fun or to progress, most of it is just useless. So most of the time when I'm playing a game with tons of text boxes I'll end up either skipping it or read it really fast and keep going
if nothing else all games should have an opt-in instant text (opt-in so the average "I'LL NEVER GO IN THE OPTIONS MENU EVERRRRR" gamer who also doesn't read the fricking manual can't then go yell at the devs when they inevitably get lost because they skipped over the plot that went "yo go fetch this item from this mysterious desert dungeon") AND/OR have it on by default in NewGame+.
Not a hot take at all
I’m charmed by that strange joke you made
calling my boss to tell him i can't come to work today cuz there's a new adef vid i have to watch 30 times in a row
Thats the best Brilliant ad I've seen.
Aside from that, damn I still don't understand how you manage to make the most boring subject ever so frickin interesting. Seeing a new ADEF video in my notifs is always a rush 😂
"Unless you want to, I don't know, I'm not your dad." In before years from now when adef's future kid stumbles upon this video and decides to attempt the run.
Also medium fast was listed on the xp curves list twice.
Not gonna lie, the Charm quark joke got a genuine laugh out of me once the required xp numbers appeared.
As someone doing their phd in physics, i laughed at the charm joke. Many thanks!
So I've been thinking that you sound like someone else I listen to, and I couldn't quite put my finger on it until just now. Your voice and inflections sound SO MUCH like stand-up comedian Drew Lynch, but without the stutter.
Also, the Charm joke was cute. I slapped my knee, even!
I think theres a fundamental misunderstanding of the word "odds" and you did a great job laying it out here
Im a particle physics grad student and I approve of the charm quark joke. The one smile I had today
thank you for the physic joke at the start
made me smile
He’s a genius
I ... I just watched a Brilliant ad at the end. The whole thing. adef is completely fired up, take care guys and watch out for an unhinged mathematician i guess. i´m not that deep into adef lore
Somehow this video turned out to be exactly what I thought it would be yesterday when I saw Zigzagoon.
Since we care about the number of rare candies received rather than whether pickup activated, the expected value is more relevant. And you can always just add expected values together, even for events that aren't independent.