I don't think HU is the most complex either. But 9 handed isn't more complex based solely on the number of players. You get diminishing returns on complexity when you add additional players.

All I said is that simply adding players doesn't make a game more complicated. 10 handed isn't more complicated than 9. The decision tree for most situations gets significantly simpler as a direct result of how additional players effect the strategy.

Most hands are players HU. Variable bet sizes can be groups.

Ugh, I think I'm done. Having actually done work on both math, and programming, I already know where this conversation is going.

Not bad at math but maybe somewhat gullible for believing you when you said such things were not that hard to calculate.

Both chess and poker have an open-ended amount of work that can be done on someone's game to get better - they are both sufficiently complex for things like professionals and world champions to exist.

Doing work on chess leads to a bigger marginal improvement in winrate than in poker. In poker even a (profitable) micro stakes full-ring player like me is probably no worse than a 65-35 dog in a HU SNG against some of the best players in the world. In chess an elderly GM Bronstein came to our club and played us simultaneously and smashed us, winning every single game and as others have posted there would be a lot of 65-35 steps between us and a current world champion.

On the other hand, doing work on poker leads to a bigger marginal improvement in finances than doing work on chess.

They aren't.

I'm just not going to waste my time proving why a stupid question is stupid. And given the example you used for how winrates work in each game just proved further why this discussion would be pointless.

Your odds of winning a single hu s&g vs a good player is not a poker equivalent to playing a good player in a game of chess.

If you used a reasonable example, you'd be 0% in both cases.


Not bad at math indeed.

Assuming 100NL 6-max (with typical rake for that level) and an equally-skilled opponent, flatting and raising have the same EV. About 0.53bb.
For HU 100NL, calling will have an EV of about 0.36bb, but raising is significantly better at 0.66bb.

I didn't do any calculations. There are various software tools that have done the math, and they are getting more and more accurate every day.
Poker is being 'solved' in much the same way as chess is. Brute force and the law of large numbers. If you simulate trillions of hands (or buy a billion hand histories for a pretty good approximation), the EV numbers converge to their long-term averages, since there's only a few decisions per hand and a finite number of board textures.

With chess, the decision trees quickly spiral out of control. After just 5 moves each, there are 4,897,256 different boards that are possible. Do you know the best sixth move on all of them?

<3

i apologise in advance if this has already been adressed ITT(i haven't finished reading):

what do those %s mean?

i assume that in chess they refer simply to probability of winning 1 game of chess?

in poker the % would vary wildly depending on what you define as '1 game'. e.g. for a single standard structure HUSNG differences in win probabilities are relatively much smaller than that, whereas imo they could be similar if you define '1 game' as a HUNL cash match of 5k hands, for example. or just change the format of the HUSNG. (i've been a HUNL reg and there would definitely be tens of such sets of players in which the best of each set wins 65% against the worst. although it would be complicated by the intransitivity of poker edge. i don't know if that applies to chess too.)

but then again my point may be irrelevant, as you may be able to make such definitions for any game/activity, simply increasing sample size to increase % of the better player winning overall. (e.g. in chess if, for example, player A is 65% to win 1 game vs player B, then that % would be way higher for a best of 3 between those 2 players)

maybe one should consider that there are many factors, e.g. stack depth, number of players, blind structure, cashgame or tournament payout, etc., that each essentially constitute a different game of different complexity etc., so one should think about what exact meaning of 'poker' (which exact format of poker) we're comparing with chess. (also, within each format of poker there are various ways to define 'winning', and thus win%s or edges)

if we're comparing a single hand(with certain stack sizes etc.) to a single game of chess, then the fact that there is inherent variance in poker but not in chess results in much lower edges in poker, and the data we get isn't simply a result (and thus an indicator) of skill difference, but rather a result of {skill difference, variance}

i brought up the issue of variance in poker vs lack of variance in chess, but it occurs to me that one thing that complicates/clouds this issue is the following:
there is some kind variance in chess too, otherwise we wouldn't be talking about 65% to win etc.
the difference is that this variance is variance in execution (we're not perfect robots that follow an exact algorythm, our mental state etc is changable and our choices are influenced by randomizing factors, and thus the *better* player may end up executing worse than *worse* player and lose), whereas in poker there is variance in execution and variance inherent in the game. (and, btw, the variance inherent in the poker hugely increases the variance in execution through tilt.)
one could look at it this way: variance in execution determines who is the better, more skillful player at that precise moment or for the duration of 1 game/sample. thus, the player who is better at that moment still wins 100% if there is no inherent variance in a game.

Using a computer to collect stats on play and calculate the best move isn't "solving" in the sense we would apply it to chess or other problems in game theory - we don't say the first move of chess is solved because the stats say e4 wins most often. It's also not the same thing as the EV being not hard for a human to calculate - which is was Lander's assertion. However it probably is the way to create computer programs that can crush even the best human players, with all the consequences that entails.

Yes, the probability of a win, with a draw counting as half a win.

Yes, it depends what you define as one game - I was talking about a HUSNG before but I was vague about stack sizes, blind levels etc. In chess the world championship is (or was) decided playing 24 matches over 24 days - because a single game of course wouldn't be enough. I have no idea but I wonder whether playing HUSNGs 8 hours per day for 24 days straight would be enough to determine a champion or if it would still be pretty much random between the best players.

On the last part. Do you mean where the worst player of one set is also the best player of the next set down? I would say I'm probably one 65-35 step down from the the Colmans and the Junglemans (depending on stack sizes), and my 7 year-old daughter is one 65-35 step down from me in the kitchen table play money game - so two steps total. The equivalent number in chess is an order of magnitude higher.

In a sense games like basketball and football are also games of chance as the "best" team doesn't always win - nor even the team that plays best on the day

I wish being ignorant was painful.

You would be in such agony.

For the record, I didnt say ev in poker isn't hard for a human to calculate.

I said it isn't hard to calculate, and that I'd like to see a human calculate ev in chess.

Do you see the differences in those two statements? Its very significant. You should work on reading comprehension and information retention before attempting game theory.

Edit: I'm aware I'm being a total dick. Its intentional. I hate when people enter topics, and proceed to talk with complete certainty on things they really know nothing about. Asking questions more, and speaking less, when you're uncertain about something is a much more intelligent action than speaking from your ass.

Though your flatulance and sphincter control is quite impressive.

In chess, that percentage is not strictly the probability of winning, because draws are possible. Counting a draw as half a point, win as one point and loss as zero, the better player should expect to score 65 points in a 100 game match. Or, for a single game, an expectancy of 0.65 for the stronger player.

But there is definitely some intransitivity in chess. And playing 100 games against someone will also allow them to learn each others style of play and openings. IMO this tends to help the weaker player - he benefits more by repeatedly playing a slightly stronger opponent than his opponent does by playing a slightly weaker one.

So it's more accurate to say that a player should score 65 points against a randomly chosen field of 100 players roughly 100 rating points below him.

Also, just because there are unpredictable variables that may or may not have an effect on a game does not make it a game of chance. IE The wind effecting a baseball game does not make it a game of chance. A chess player missing something they often wouldn't isn't "variance".

Randomness and things appearing to be random are vastly different things.

So the correct comprehension of

is not making a comparison as part of a chess and poker discussion but as two separate unrelated statements nicely separated into two paragraphs? Ok.

Really? Lol

You just keep lowering the bar for yourself

Let me rephrase.

There is a comparison there, you're just not understanding it, and that's causing you to grossly misrepresent what I said.

Humans are able to calculate EV in poker with some degree of accuracy, though a computer is necessary for real precision. However, this is true for most things.

Even the #1 chess player in the world can't calculate the EV of a chess position with any degree of accuracy, and it takes an extremely sophisticated engine to actually do so. The best players in the world will have a sense of whether or not a position is winning, losing or equal. A human can't even begin to quantify what the EV in a chess position is in anything but the most vague terms.

Want to take a guess as to why that is?

Why bother, I'm sure you're going to misunderstand it and respond with something wildly inaccurate, as you have with nearly every single post.

I guess I'll have to keep my opinions to myself since apparently I know nothing about the complexity of poker or chess.

I haven't played on icc in years. My sn on chess.com is the same as the one on here.

I'll play a game of chsss anytime

Agreed, but you don't have to be a world class player to be able to say - on move 20, for example - something like "This position looks slightly better for white, because of black's bad bishop". But a program like Fritz can analyse the situation and calculate that white is "winning" by 0.2 pawns in relative piece value, because of the positional advantage. Converting the slight positional advantage into a win still requires a lot of skill, of course, and a bad move 21 could actually turn the advantage in black's favour.
In poker, the situation is often much clearer. If you have top set aces on the turn and there's no flush or straight draw, villain's entire range is drawing dead, so no skill is required to win the hand, and maximising EV will likely come down to two simple decisions for the player with the nuts, neither of which are "fold".

FWIW, I do agree that there is a lot of complexity in NLH, but I don't think it's anywhere near as complex as chess.

That is not true at all. Even I can give a relatively accurate estimate of most positions as long as there aren't crazy complications... Are you referring to EV in the sense given a position and the particular players the chance that one will be able to convert the position into a win?

Also Lander do you even play chess? Given some of your post it seems you know nothing about the game.

We aren't talking about chess, now are we? We're talking about complexity in poker and chess, and which is more complicated.

A "relatively accurate estimate of most positions" is not the same as "I can tell you the EV of this situation."

The computer can tell you the EV. It takes a degree of precision that humans aren't capable of in chess. Just because the answers might mean virtually the same thing, it's not the same. A grandmaster saying "mmm...I think this position is winning and white has a slight advantage" might mean close to the same thing, but it's not saying you have a 0.86 pawn advantage.

A human of a similar skill level(and even significantly lower) in poker is much closer to achieving that degree of precision, because quite frankly, it's not that complicated in comparison. Obviously there are exceptions, as this is less true for Omaha than in Hold Em, and further still in 5 card stud.

I imagine most people are talking about hold em, a game with only 1326 starting hands. Of these, there are only 169 unique starting hands. Most of which will be dropped at the beginning of your calculation, and will be narrowed down even further as the hand proceeds(which is why 9 players isn't suddenly more complicated). More often than not, but not always, your decision is between two choices(bet or check, call or fold). And quite often, one option is enough better than the other that it's pretty clearly one way over another. Bet sizing is going to be the most complicated part of the whole process.

The numbers just aren't large enough to be outside the grasp of a human. It's by no means easy in the sense that anybody can learn it as if they were tying their shoes. However, it is a drop compared to the numbers involved in calculating a chess position with accuracy. Chess dwarfs Omaha, a game with 270k starting hands.

It's the size of the numbers involved that make it impossible for even a grandmaster to tell the EV of even very clear positions. If white is completely won short of a massive blunder, great. What's the EV of the position?

Kind of a silly question to ask me what the chance somebody can win in a position of chess, isn't it? No, I am not talking about chance in chess, a game of skill without an element of chance.

I don't believe I've said a word about the actual game of chess in this thread, unless you count the part where I talked about choosing openings that your opponent lacked experience in. That is a statement I could have made without even knowing the rules of chess, so who really knows.

Correct me if I'm wrong.

For the record, I don't actually care whether or not a grandmaster can tell me the exact EV of a position of chess. It's not particularly relevant to what he's doing, and certainly isn't some kind of slight to grandmasters. It's just a statement of how complicated the game actually is.

If a human was capable of calculating EV in chess and using it to determine the best decision, they'd already be doing it.

While I am not a professional chess player by no means I would consider myself a strong amateur. My USCF is around 2400 and I can even give a rough estimate of positions such as .25 or .31 with some degree of accuracy. Given how often chess engines are used to train these days Grandmasters could easily do better then ..... Ohhh I think this is winning. Also at a certain point the evaluation doesn't really mean anything. No real difference between +3 and +10 most of the time. While I agree that chess is more complicated then poker its really not as easy to compare the two as you are making out. Humans find the best move in chess much differently then computers do. In fact strong players in chess calculate less variations then weaker players because they eliminate many candidate moves off the bat that amateurs do not. Some moves in chess are the same thing as 72o UTG. How else could strong players play pretty accurately in 1 minute chess games if it was always based on calculating so many variations? Also I don't think chance to win is something silly. When I say chance to win it doesn't mean I am talking about luck. If a sports book estimates the chance of a team beating another team at 3-1 for example they are making a prediction based on the skill of the teams. So unless you consider something like football or basketball a game of chance I see nothing wrong with my statement.

Even Carlsen won't play "pretty accurately" in 1 minute games, but he does play very strong. And yes, they DO calculate in 1 minute games, but they know what to calculate and what not to.

You've completely missed the point, so I've lost interest. Try not to derail things, I've no interest in arguing with people on the internet.

You do realize by being so off the mark, you completely kill any motivation to have a productive discussion with you? Gives the feeling that any exchange will result in the same same feeling of talking to a brick wall.

I liked the part where somebody making a prediction is sinilar to somebody's chance of winning. Especially when talking about comparing team games to heads up games.

Oh, and the part where strong players calculate less variations than weak players because they have fewer starting points. Fewer candidate moves is not the same as fewer variations.


"Rough estimate of positions" and "some accuracy" was cute too. Especially as I'm talking about precision.


I'm not responding to you unless your next post isnt stupid. I'd say chances aren't good though. I'd put my EV at 0.07 pwns...