Chess AINEC.

I read a game theory whiz's comments on this once, and I think he said the game space of chess is a lot larger than the game space of poker. The reason computers have made more progress on chess than poker is vastly more resources have been devoted to solving chess than to solving the various forms of poker.

But that is the strictly theoretical answer to the question. More importantly, *in practice* poker play is vastly less complicated than chess play.

As bad as the human mind is at charting out the consequence of a chain of possible moves, which is the essence of the chess thought process, it's even worse at keeping in memory an array of weighted probabilities, then splitting these probabilities into further probabilities conditioned by different boards and different actions, then eventually summing them up and balancing them in a way that is optimal across the entire range of possible opponent counterstrategies.

People can play chess pretty close to optimally. They will never be able to play non short-stack poker close to optimally without the aid of a computer. (BTW by "close to optimal" here I mean the frequency with which the human diverges from GTO strategy, not the EV that the human has versus a GTO computer...most of the divergences in poker probably won't cost much, if anything at all.)

There's a further complication from the fact that high limit poker players are much worse at poker than high level chess players are at chess. People play poker to gamble, people play chess to win. That's why poker is vastly more profitable than chess.

Therefore it's not necessary, and in many cases not even desirable, to make great efforts to play GTO.

And for the record most poker players who claim to play a largely game-theoretical style of poker are full of ****. That includes the LHE players who ran really hot for a while.

why bother reading a game theory whiz's thoughts when you seem to be such an authority yourself!

Meh, not sure where the sarcasm is coming from.

Reading your other comments in this thread you waste a lot of energy on the 9-handed versus 2-handed versus 9-handed-folded-to-2 thing.

It goes without saying in a discussion like this that we are talking about HU poker. There is (probably? definitely? not sure) no solution to multi-handed poker because of the possibility of unintentionally collusive strategies where one person does stupid things that hurt another person but benefit the third.

As for whether HU poker is solvable, it is according to Nash's theorem. The reason it hasn't been is this is a huge computational task and nobody has devoted the necessary resources to the problem.

Wait, what? When did we go with this without saying it?

I think that first you got to refine what "complex" is.

I don't mean OP's general undefined topic, I mean the specific question of which game is more difficult to solve.

Since multiway poker is unsolvable obviously any discussion about the solution to poker is implicitly referring to HU poker.

multiway poker is obviously solvable, just it's way too complicated for us to solve right now.

Wow, this is something I had not considered and I agree that it seems to really confuse the possibility of solving poker. If three players are all playing optimally and then suddenly two of them have access to any sort of additional information then that alone would make them winners and the first player a loser even though he would be playing "optimally".

It doesn't even have to be collusion but any sort of information that the first player simply is not privy too.

Fun question,
Its one I have argued on many occasions with some of my poker friends.
I come from a chess background, and have played the game in competitive events on many occasion, just like I have poker. As of late I have dedicated way more time to poker, as it has taken over as my favorite game.

However I think there is two ways of answering this question;

1. What do u consider to be complicated? Maybe you have a different way of explaining "complicated" than its original definition.

2. Literal explanation.

Complicated
–adjective
1. composed of elaborately interconnected parts; complex: complicated apparatus for measuring brain functions.
2. difficult to analyze, understand, explain, etc.: a complicated problem.

Well I see most poker players that dont really know much about pokers history say that all moves and variations have been figured out in poker. So someone who has been able to learn all of these variants of the opening game, mid game, and end game will always win.

Well I will partly agree, thats correct. If you have the capacity to learn all of these stages, and different openings, than yes, you in theory could begin to play "abc chess" and proceed to win, the problem with this statement is it will take you a life time.
The game of chess has been studied for hundreds of yrs and its still being studied. Why do people have different openings? defenses? well because just like in poker playing TAG, LAG and so forth has its own benefits vs certain opponents or situations.
Now if you still want to discredit that statement, than look a the Magician of chess Mikhail Tal this man was literally a world champion and one of the biggest parts of his game was bluffs, yup u heard it right bluffs, he would literally use inferior play, or what would normally be considered wrong by a chess pundit to get into his opponents head and psych him out by maybe sacrificing a queen in the middle game to gain a mental advantage over his opponent.
Now I could go on with the depth of that kind of play, but ill move on

Now, poker.
There maybe is a bigger stress level to poker than chess as you stand to lose you life savings in 10min but that dosent make it more complicated.
I mean I remember my first ever sit n go, I never played hold em in my life when i decided to sign up to a game online and got 8th place in a 400ppl tournament.
This kind of result would never be attained in a chess tournament, if you have never played the game.
You could argue I got lucky, and I would agree, but the point is the game of poker can be learned in a matter of minutes, and level of proficiency can be attained rather quickly as well. This can not be done in chess.

Final argument, for simplicity's sake...

But than again all games are complicated, and it depends on what your good at, I mean im sure I maybe can play connect 4 better than "Bob", but Bob can smoke me at Tetris. So it could be argued that all games are complicated at some degree and to some people. But if you were to take a 100 random strangers that never played the poker and chess and have them coached by the best pros in the world (in each game) for 24 hours and than take them and make them play their coach the next day, im sure the poker coach would lose way more games out of 100, than the chess coach, in fact im quite sure the chess coach would lose 0 games


Cheers

this is a massive copout - you are essentially saying it will be easier to solve because (among other things) our opponents will play much worse in some situations. You may as well say "playing go against 6 year olds is solved".

this point is valid so long as you are talking about beating a game, not "solving" it or "playing perfectly" or "exploiting your opponents maximally". But that's a red herring. I already said it might be easier to program a bot to beat a typical 9-handed game, but much less complex to solve the HU game.

It seems like in every post you are using a different definition of "solve".

That's not what the point of what I was saying was. In a full ring game the hands folded to you on the button makes up less than 5% of your hands in the absolute best case (10% vpip of every single other player in the game, with nobody else willing to exploit the tightness of the game so the math is simple: you get the button 10% of the time in a 10-handed game, it's folded to the button .9^7 = 48% of the time so you get the situation .1 * .48 = .048 = 4.8% at the absolute max).

In a 3-handed game, you obviously get this situation 33% of the time, and I think it is clearly one of the most difficult to solve in and of itself especially against opponents who are more likely to be playing near optimally themselves as well.

The point is that if that if you don't manage to solve this situation in a full-ring game then it won't be too huge a deal - as you have plenty of other situations that occur much more frequently and are vastly easier to solve. Eg - one of the 10% players opens in early or middle and you have 77, you have JQ, you have KK, etc. And it isn't just some random aside or cop-out, that in a 9-handed game not only will this situation occur vastly less frequently, but in reality it will also be substantially easier to reach near perfect play than in the 3-handed game.

Basically, my emphasis is on the process of solving. What I am saying is that to solve 90% (such that you are playing perfectly in 90% of situations) of full ring would be vastly easier than to solve 90% of 3-handed.

So I suppose a cliff notes version would be that I would expect it would take 10 units of time to solve full ring for 95% of decisions. Whereas it might take something like 100 units of time to solve 3-handed for 95% of decisions. But then as 3-handed is technically a subset of 9-handed then if the time to completely solve 3-handed would be 110 units, then the complete time to solve full ring would be 110 + 10.

So I guess I agree that full ring would be more challenging to solve as a whole. But in practical terms the process of solving it is vastly easier.

So I suppose that brings up an interesting question. Which game is more complex as a whole. The one where 90% of decisions are trivial and basic, but 10% have the potential to be incredibly complex. Or the game where 100% of decisions are extremely complex, but the most complex decisions must technically be slightly less complex than the most complex decisions of the other game?

I still think it is clearly the latter.

that's not "solving it", that's just "playing well".

It's trivial to have a computer program that "solves" 90% of chess positions, in fact any decent program can probably solve 99%+ within a fraction of a second each. To solve a game, whether you want to play unexploitably or to make the max profit, you need to solve the interesting cass -- the other 1%.

Oh come on. That's obviously wrong and I think you know it. A position in chess is absolutely NOT even remotely analogous to a round of poker. A round of poker is analogous to an entire game. If there was a chess program that played perfectly in 99% of chess games it would be the strongest program ever, by a landslide. And similarly if a program could play perfect chess in 99% of chess rounds/etc then it would likely be the strongest/most profitable/etc poker player even if it just folded the other 1% regardless.

A single decision in poker is analogous to a single decision in chess. Of course there are more choices in the chess situation, if you don't count trivial things like "bet $20" vs "bet $20.01" but it's the closest way to compare them.

Again, no it's not at all.

If you played very well in 99% of chess positions and mediocre 1% then you're going to be quite mediocre at chess overall - that's nearly .5 blunders per game on average. If you play very well in 99% of poker decisions and mediocre in 1% then you're still going to be incredibly profitable and almost certainly one of [if not] the best in the world.

this.


good luck becoming a grand master if you learn the game at age 18.

Since when does playing mediocre = blundering. Sounds more like making a suboptimal move every now and then, which everyone does probably more than 1% of the time.

I'm pretty sure it's considered by game theory people not to be solvable. The reason is that for any strategy player A picks, there are deviations from the seemingly optimal strategy that B & C can make in tandem that will decrease A's expected value.

For example imagine a 3-handed LHE game where everyone is playing what we would imagine to be "optimal".

Now imagine that player A, when and only when he has the button, decides to open 100% in this 3-handed game. which is clearly not "optimal".

Now say player B when and only when he is in the SB decides to 3-bet 100% of the time, again clearly not "optimal".

Now player C in the BB, playing what he previously though was the optimal strategy, will be folding over 90% of the time by my estimate. If less, not way less.

Yeah he will win bigger pots than he would have one against optimal players when he does play, but assuming A/B play reasonably postflop this scenario represents, imo, a clear loss of EV for player C.

---

Obviously someone from the peanut gallery will say that player C should just jack up his calling and 4-betting frequencies when he is in the BB, but obviously that new strategy becomes exploitable if players A and B decide to cool down.

The bottom line is in multiway poker there is no strategy against which your opponents can't improve their EV by collusive exploitation (and no this doesn't mean "collusion" or cheating.)

Ok I'm no big expert on the topic. Instinctively seems like it should be solvable but if people who actually know something say it's not then who am I to argue.

It sort of makes sense what you are saying...so basically there is no game theoretically optimal choice that wins or at least breaks even regardless of what your opponent's method of play is?

I believe that's what the game theory experts assert, for multiway poker.

I think you're probably right (though obviously we can't know for sure)

can you describe a simple game where this is true? I can't figure out a good example.

But HU poker (even NL) is theoretically solvable, right? I didn't know that game theorists thought that multiway poker was unsolvable. In my opinion, if they hold such a strong opinion there must be a proof, and if there is a proof, then a post in a forum referring to this opinion supposedly held by game theorists should have a citation.

I was trying to give him the benefit of the doubt since if you change it to just a mediocre decision then it skews it even more. If you play very well 99% of the time in poker and only make a mediocre decision 1% of the time then you would be by far the best player the world has [or likely ever will] seen. I have no idea what it would mean in chess. I guess the exact definition and range of mediocrity becomes critical.

Point being all analogies work and balance perfectly when you use the proper analogy of chess game to poker round.