Hacker News Reader

Comment by HPsquared | Hacker News Reader

Chess is logic, poker is rhetoric.

This is not true in the slightest.

Poker is far, far more mathematical than chess. The idea that poker pros are only going off "reads" is completely ludicrous.

They have a set game-theory-optimal style of playing, and then use an opponents flaws/"tells" to deviate from the GTO style to exploit those tells.

An example of a flaw could be if the opponent is playing shitty hole cards.

If the opponent has played in every pot over the last 30 minutes (something you'll frequently see), then he's playing way too many hands preflop.

This is exploitable.

Just a super basic example, but you'd be surprised at how profitable this is. Beginners play way too many hands. If you want to make money, just fold every hand and let the beginner collect the blinds until you get AAs, KKs, QQs, etc. and then you should bet big. The beginner won't notice and he'll pay you off anyway. You'll be massively +EV.

A more advanced player will notice what you're doing and he'll immediately fold when you bet big. Therefore, you'll have to start incorporating "bluffs" into your strategy so his tactic of folding whenever you bet big becomes unprofitable.

Poker is about figuring out all of these things and then using these tools to make the most money.

Bill Chen wrote a great book called "The Mathematics of Poker" that goes through all the math involved.

It's all probability theory and game theory.

He gave a talk at MIT that goes over some of the basic concepts -> https://youtu.be/BuxCNZ0RVKA

The obvious issue with teaching your kids poker is that it'll be very difficult for your kid to find anyone else to play with. He/she can't go play poker in a casino obviously

The strategy you suggest seems too simplistic and predictable to even trick a beginner to be honest

It might seem to be to simple, but don’t underestimate the power of boredom and curiosity. The new player won’t know what you’ve been folding the whole time and will have had nothing else to do but call a sudden change in behavior to see what you have, or fold and keep the pattern going.

One player (beginner) wants to have fun (maybe not overtly, but this is usually the case), the other wants to make money.

I’ve used the strategy in some pretty big games. People get bored, drunk, lazy, etc. It’s pretty easy to consistently make money at poker if you aren’t playing anyone who takes it seriously.

You’d be surprised how boring it gets at a poker table when you don’t get good cards for a hour. Folks get bored and play suboptimally.

Part of what poker teaches is patience.

Yep, boredom is why I quit playing poker in casinos. I'm usually at a casino with friends to have a good time, but winning poker requires folding the large majority of hands [1], and it's just not fun. Craps with friends on the other hand, but I digress...

Also, very different than small poker games with friends where the point is hanging out and 'gambling' a bit.

[1] Tournament style is different from an open ended game.

This is why David Sirlin made a poker variant which is fun to play at the dining room table

https://www.sirlin.net/articles/designing-pandante

And he did so by making more hands good so you don't want to fold all the time

No, total beginners have all their mental power focused on trying to find patterns between the cards on the table and the ones they have on their hands, while trying to remember all the combinations of poker hands. They also need to pay attention on which turn is next and how much they need to pay to keep playing and figure out if it’s worth it, they also need to learn the common poker language, a lot of things going on for a beginner. Once you have all that in memory it becomes easier, your processor power is free from those memory tasks and you can start reading people’s intentions.

The fancy deep learning poker bots that beat pros don’t even consider past opponent behavior.

If you're talking about Libratus (the CMU AI that beat poker pros in heads up no limit), then that's not true I believe.

During the competition, Libratus used 4 million core hours on the Bridges supercomputer for analyzing it's prior gameplay and changing strategy.

> During the tournament, Libratus was competing against the players during the days. Overnight it was perfecting its strategy on its own by analysing the prior gameplay and results of the day, particularly its losses. Therefore, it was able to continuously straighten out the imperfections that the human team had discovered in their extensive analysis, resulting in a permanent arms race between the humans and Libratus.

From https://en.wikipedia.org/wiki/Libratus

If the bot is playing a Nash equilibrium then it doesn't need to devise the opponent's strategy. It's guaranteed to tie the opponent in the worst case (the opponent is also playing Nash equilibrium) or else win.

https://en.wikipedia.org/wiki/Nash_equilibrium

However, if the opponent is playing sub-optimally you can win _more_ by deviating from Nash equilibrium to take advantage of their specific strategic shortcomings. The risk in doing so is if you are wrong estimating their shortcomings.. you are no longer guaranteed a tie in the worst case scenario.

You can see this play out with a small toy game and a card calculator. Suppose a two player game of Texas Holdem where each player has 8x big blinds. The first player must choose to either go all in or fold. The second player must choose to either call the all in or fold. Players must pick a range of hands to perform either action prior to looking at their cards.

https://openpokertools.com/range_equity.html

Suppose a silly strategy by the second player of folding any hand except pocket aces. You'd be wise as the first player to go all in with any hand and pick up a free big blind with 99.5% certainty. However going all in with every hand is also an easy strategy to take advantage of...

(You can actually go back and forth maximally exploiting the other player's strategy and you will eventually reach Nash equilibrium)

On its face that seems impossible to be true. Given any bet you could calculate the probability that the bot would fold given an all in raise if you had perfect knowledge of the bot and abuse it

If you can abuse knowledge of the bot's strategy then the bot is by definition not playing a game theory optimal strategy.

At some point I want to write a long-form post with code proving my statements above. I walked through this scenario a long time ago using https://www.flopzilla.com/holdeq.html to find the equilibrium but it'd be neat to write open source code that does the number crunching and that way others could actually audit the proof.

The short of it though is there exists a perfect range of cards the bot can play where it neither folds enough for you to bluff to gain an edge, nor does it call enough that you can wait for better cards while bleeding chips to gain an edge.

The reason I think this isn’t theoretically possible is that the bets aren’t binary. You have to specify a number. So with knowledge of the bot you could probably set up some strong priors on what they have.

I say this despite also claiming the top DL bot does this anyway and does very well.

Definitely not true for the top bots.

Any normal chess player can beat the best poker player in the world with a little luck.

No poker player can beat the best chess player in the world unless they aren't at their level already.

I believe over multiple poker hands that advantage from luck would quickly diminish and then disappear. I would be very surprised if any normal chess player, a total novice at poker, would be able to beat the best poker player in the world if they sat down to play for a couple of hours straight.

Variance in poker evens out over months, not a single session. A competent but not great player can definitely beat a pro in a single session.

I think you're overstating the variance. If you watch high level poker play you will consistently see the same players at the final table of big tournaments. Sure, a few lucky random players will go pretty deep in the tournament, but they just don't do it nearly as frequently as the top few guys.

A competent but not great player has a chance of beating a pro in one session, but it's definitely pretty slim.

When poker players measure variance, it's $$$ over time. And while some pros are indeed good are surviving the bubble in fishy tournaments, none make a perfectly consistent income. Even amongst crushers, the variance can be massive.

For every tournament you make a big score, you crash out of 5 and get min payout on a few more.

Many of the biggest names in poker have been the biggest losers in a given calendar period. For example, Gus Hansen and Daniel Negreanu are two of the biggest winners of all time yet also have some of the biggest documented losses...

Not clear how long your "sessions" are but they have poker tournaments where the same people regularly make it to the top.

Not as regularly as you think. Earnings vary a lot.

As for how long a session is? 8 hours. I made a living from poker for a few years.

Skill would win out, but "hours" is not even the right ballpark.

For comparison, 50K hands is a standard length for heads up matches between pros, with the understanding that even then the best player is not guaranteed to win. In live play, you'd be looking at 50-100 hands per hour.

It's true that with a greater skill gap you'd need less time, but you'd still be looking at thousands of hands if you wanted something like 95% confidence that the better player was up.

It's a bit like how the Spartans (IIRC) taught their kids to lie, cheat, and steal as long as they didn't get caught, because that's how the world works. Unfortunately, this kind of society seems doomed to collapse.

This strategy only works if most people don't expect you to lie, cheat and steal. Once the majority does it, you end up with the massive overhead of a trustless society while not gaining anything.

Reply on news.ycombinator.com