Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why Can a Machine Beat Mario but Not Pokemon?

medium.com

Post reply on HN
Print view View original

41–50 of 76 posts

Re: Why Can a Machine Beat Mario but Not Pokemon?

#41

Post by Aissen »

andrewmcwatters wrote: post #37
Aissen wrote: post #4

Earlier quoted context omitted.

If you're following video games, Pokémon isn't really anything special. It's an RPG. It's doesn't have any innovative gameplay. What was new at release, is the way it combines things which already existed at the time. For example, in Final Fantasy V (which came out 3 years before pokémon), a character could already catch monsters and have them battle against others. Many games already had a collecting element. The ex…

> If you're following video games, Pokémon isn't really anything special. It's an RPG. Pokemon Red, Green, and Blue are the number two Game Boy video games of all time. Behind Tetris of all games. This is a pretty extreme reductive statement, akin to calling Super Mario 64 "just a 3D platformer."

I think you missed the point. I'm not talking about popularity, but game mechanics, see my other comment.

Re: Why Can a Machine Beat Mario but Not Pokemon?

#42

Post by catwell »

varunramesh wrote: post #28
CM30 wrote: post #9

I think the battling issue may be a hard one for AI to deal with alone, though not so much in terms of the main game. Remember, the main story is meant to be beatable by people who don't want to think very much and who don't really care about training their team beyond 'reach a certain level'. Hence in the main story, 'grind until you significantly outlevel the next opponent' would probably be an optimum strategy for…

In 2014, I wrote a minimax-based AI to play games on Pokemon Showdown. We adapted Showdown's battle simulator for our tree search. The hardest part was syncing the local simulator state with the actual game state - the battle state in Pokemon is both complex and partially observed. Bugs in this process could result in the AI using Protect twice because the state wasn't updated with the fact that it used Protect the p…

This is super interesting. I would like to see a real (learning, à la AlphaGo) A.I. play on Showdown OU someday, without cheating (i.e. about the number of matches per day a real human player would play).

I think there are a few challenges not found in most other online games, one being that strategies that win on different strata of the ladder are not the same (e.g. hyper offense is the most efficient on the lower ladder, but at some point around 1500 / 1600 you will start losing using it...). Also, I wonder how well an A.I. trained on the ladder would do in a tournament (e.g. SPL), where the metagame is a bit different.

Your code looks like a good entry point for that, all that's needed is to write a new bot using ML libraries...

Re: Why Can a Machine Beat Mario but Not Pokemon?

#43

Post by Retric »

chongli wrote: post #13
chongli wrote: post #8

Earlier quoted context omitted.

All of the details you're referring to are way above the level of any machine that plays Mario. For SMB1 the branching factor is uncommonly low. Pokemon has an extremely high branching factor by comparison. This being measured at the level of controller inputs and memory content outputs. SMB1 can be beaten by applying random controller inputs, measuring the player's x value, and seeing what works. Yes, you do need to…

To be fair to the machine, a human wouldn't even be able to tell what game they're playing, let alone be able to finish it, if all he had to go on was the raw memory values in response to controller inputs. The fact that Mario can be beaten at all given such an incredible dearth of information is impressive. RPGs and action adventure games like Pokemon or Zelda are too laden with human-specific cultural details: hero…

I once watched my 8 year old sister play an RPG in a language she did not understand. She had only the vaguest idea what was going on but still made progress. Just randomly trying things looking for novelty is enough to progress through such games given sufficient time.

Re: Why Can a Machine Beat Mario but Not Pokemon?

#44

Post by jknightco »

ziaddotcom wrote: post #35
kd0amg wrote: post #22

Earlier quoted context omitted.

A pokemon can be one (or two) of a bunch of different types like "bug" or "rock" or "ghost," each of which comes with a particular set of types it's vulnerable to (2x damage taken) and another set of types it's resistant to (0.5x or 0 damage taken). So no one pokemon has an advantage over all the others, and there's no 2-3-4-...-J-Q-K-A strength ordering for types. You can also freely substitute among your team of 6,…

Right, but IRL blackjack the human is the pokemon, vulerable to various type of human tells etc to what is otherwise grindable. As far as the various pokemon vs another vs the simplicity of jkqa it seems to me something like a weighted adjacency matrix could handle the increased complexity even in a relatively simple ai (imagine two pokemon doing a graph traversal to opposite ends of a node graph, whichever gets ther…

There are no "human tells" in Blackjack. Its you against the deck. Maybe you're confusing it with poker?

Re: Why Can a Machine Beat Mario but Not Pokemon?

#45

Post by esrauch »

666lumberjack wrote: post #23
ziaddotcom wrote: post #16

Earlier quoted context omitted.

Is the battle system much different than playing blackjack with your own deck? Imagine a world where someone can buy and trade towards a deck with 90% aces going up a person with 80% aces.

Most people who battle at a high level play on a simulator where any legal Pokemon is available to everyone, or alternatively use tools to 'hack' whatever Pokemon they want into the game. Some of the most popular strategies use Pokemon/Move combinations that were only available through events 10+ years ago, but since it's possible to transfer those Pokemon forward to the latest games those moves are available to use.…

They can't possibly verify that the Pokemon came from a network of breeders instead of one of them hacking it in right? Is it just that if you have a suspiciously perfect lineup you need at least a plausible sourcing explanation?

Re: Why Can a Machine Beat Mario but Not Pokemon?

#46

Post by chongli »

Retric wrote: post #43
chongli wrote: post #13

Earlier quoted context omitted.

To be fair to the machine, a human wouldn't even be able to tell what game they're playing, let alone be able to finish it, if all he had to go on was the raw memory values in response to controller inputs. The fact that Mario can be beaten at all given such an incredible dearth of information is impressive. RPGs and action adventure games like Pokemon or Zelda are too laden with human-specific cultural details: hero…

I once watched my 8 year old sister play an RPG in a language she did not understand. She had only the vaguest idea what was going on but still made progress. Just randomly trying things looking for novelty is enough to progress through such games given sufficient time.

The RPG may have been in another language, but it presumably still depicted human or humanoid characters, objects, enemies, etc.

Your sister knew far more about life (and consequently the game) than what a computer would know. Just having a basic concept of reality, that the world is made up of objects and agents, that some things can be interacted with, cause and effect, etc. is way beyond what a computer has (which is basically nothing).

To a computer playing a game, "randomly trying things" has only one meaning: random control inputs. As mentioned earlier, random control inputs will not get you very far in any open world game. A computer playing Zelda would be lucky to reach the bottom right corner of the map (by maximizing x and y positions) using only random inputs and reinforcement learning. Actually completing the game is so far beyond that it's ridiculous.

Re: Why Can a Machine Beat Mario but Not Pokemon?

#47

Post by WorkLifeBalance »

setr wrote: post #20
CM30 wrote: post #9

I think the battling issue may be a hard one for AI to deal with alone, though not so much in terms of the main game. Remember, the main story is meant to be beatable by people who don't want to think very much and who don't really care about training their team beyond 'reach a certain level'. Hence in the main story, 'grind until you significantly outlevel the next opponent' would probably be an optimum strategy for…

I think one major difference between pvp games and games like mario/pokemon campaign is that in the former case, the machine is competing with a meat learning algorithm. That is, it doesn’t have to find a good enough solution to a static map, but best an alternative optimization strategy (powered by wetware) before it can reasonably claim success. That, and ofc, games with a competitive scene come about because of th…

Meta doesn't imply cycling, just that the best strategy may be a mixed strategy which involves randomly picking between different pure strategies.

Instead of cycling, the meta ought to converge to a Nash equilibrium.

With the right mixed strategy, an opponent choosing a pure strategy would be at a disadvantage.

Randomising over a huge choice of pure strategies may be infeasible of course and in the real world players have to train for particular strategies which is why we see meta shifts. (Plus of course in the real world the conditions (assumptions) change due to gameplay patches.

Re: Why Can a Machine Beat Mario but Not Pokemon?

#48

Post by 666lumberjack »

esrauch wrote: post #45
666lumberjack wrote: post #23

Earlier quoted context omitted.

Most people who battle at a high level play on a simulator where any legal Pokemon is available to everyone, or alternatively use tools to 'hack' whatever Pokemon they want into the game. Some of the most popular strategies use Pokemon/Move combinations that were only available through events 10+ years ago, but since it's possible to transfer those Pokemon forward to the latest games those moves are available to use.…

They can't possibly verify that the Pokemon came from a network of breeders instead of one of them hacking it in right? Is it just that if you have a suspiciously perfect lineup you need at least a plausible sourcing explanation?

There are a surprising number of hidden values / details with which a naive hacker could be tripped up - especially for event Pokemon - but it is definitely possible to create undetectable hacked Pokemon with enough work. There are also cloning glitches in many of the games that allow one perfect Pokemon to be replicated and distributed to multiple people, although that's mostly used for casual play because it'd be suspicious for multiple competitors to show up at a tournament with exactly identical Pokemon.

Re: Why Can a Machine Beat Mario but Not Pokemon?

#49

Post by VLM »

jknightco wrote: post #44
ziaddotcom wrote: post #35

Earlier quoted context omitted.

Right, but IRL blackjack the human is the pokemon, vulerable to various type of human tells etc to what is otherwise grindable. As far as the various pokemon vs another vs the simplicity of jkqa it seems to me something like a weighted adjacency matrix could handle the increased complexity even in a relatively simple ai (imagine two pokemon doing a graph traversal to opposite ends of a node graph, whichever gets ther…

There are no "human tells" in Blackjack. Its you against the deck. Maybe you're confusing it with poker?

Any multi player game with hidden data lets you troll the other players. So if its the dealer and two or more players sharing a deck you can mess with them.

Everyone counts cards so under extremely rare circumstances given card count X you can play as if the count is X+1 or X-1 thus mess with the minds of people who accurately believe the count to be X, possibly leading to an advantage, if, statistically, they're in a position where being one off in the count is worse for them than it is for you, or if you're in such a bad spot that it doesn't matter if you signal something false you're doomed anyway.

Most of the people claiming to shuffle track, are not. Still, you got three guys watching the dealer shuffle by hand, then two can troll the third by pretending and the third's all like "what did I miss?" maybe panic and do something dumb.

Its not as much fun as poker, no.

Re: Why Can a Machine Beat Mario but Not Pokemon?

#50

Post by ghaff »

VLM wrote: post #49
jknightco wrote: post #44

Earlier quoted context omitted.

There are no "human tells" in Blackjack. Its you against the deck. Maybe you're confusing it with poker?

Any multi player game with hidden data lets you troll the other players. So if its the dealer and two or more players sharing a deck you can mess with them. Everyone counts cards so under extremely rare circumstances given card count X you can play as if the count is X+1 or X-1 thus mess with the minds of people who accurately believe the count to be X, possibly leading to an advantage, if, statistically, they're in…

But there are effectively no other players as far as you're concerned. You're playing against the dealer/house who makes purely rule-based plays.
Post reply on HN