Tuesday, May 23, 2006

Squeezing the EV Out of a Loss

I was posting a comment on Fluxer's site when, as often occurs to me, it got rather voluminous. Rather than run on in his comments, I decided to make a virtue out of verbosity and flip it into my own post.

Fluxer responded to TripJax's first question (What is the biggest mistake people make at a NL table?) as follows:

People don't realize that they will lose money after making good plays. I used to think that if my stack decreased playing a NL cash game, I was playing bad. Usually no one has a big hand, and the cards are irrelevant. Good players take advantage of that. You have to be willing to lose if you expect to win, and that's why bankroll is one of the most important aspects of poker in general, since the best players will experience the biggest swings. Poker is an exciting game because anything can happen, even if your play is consistently good or bad.

In general, I think I agree with this statement. Nothing original in the thought that one needs to focus on decisions rather than results to be successful at this often cruel game.

There is a risk, however, in relying on this way of thinking as a crutch. For one thing, it does not explicitly state the corollary point that people also forget that they will often win money after making bad plays. (Disclaimer, I'm not accusing Flux of thinking like this, I just am riffing on the idea.)

You start with a hypothesis that a certain decision is a good decision. You make that decision. You end up with a bad result. Conclusion? You're far from ready for one. Even still, if your first instinct is to challenge your own hypothesis, that is not something you should suppress. You should take it as a instance against your hypothesis, and consider why it failed.

It is absolutely true that good decisions may not cause good results and poor decisions may not cause poor results. Nevertheless, it is true that good decisions positively correlate with good results and that poor decisions positively correlate with poor results.

One must be careful before dismissing a bad result as anomalous out of hand. You need to think it through.

Sometimes, this is easy: if your move resulted in all the money going in when you were ahead, and opponent nevertheless got lucky, your hypothesis should remain unqualified (here I am assuming you are playing a ring game).

Other times, not so much. For example, say you lose when you opponent calls when he should have folded because your bets clearly represented a range of hands that were way ahead of what he called you with and even when he called you had ten outs twice, but you missed anyway. The correct treatment of this case?

The answer cannot be: you made the right decision and had bad luck. It was not bad luck that made your opponent call you. It was perhaps a number of factors: 1) it could be your table image was not tight enough to sell the hand you were representing, 2) it could be that the guy was multi-tabling and had not noticed you had folded 94% of your hands in the last 3 orbits, 3) it could be that the guy is a flaming moron. (It could also be that you got outplayed.)

Nevertheless, the natural response is going to be to type bloody murder into the chat box "how could you call that you #&$)*#$fish?!"

Mind you I'm quite guilty of this reaction myself, but I recognize it as an incorrect and worse a counterproductive response (this is a version CJ's critique of people who complain that the other guy calls him with the better hand, which may or may not have been aimed at me but I'll own it because I know I've done it). It's wrong not so much because it teaches your opponent not to make that "mistake"--they almost always believe they made a savvy call and may well enjoy your ire--but because it deflects attention from where you need it, that is, from adjusting your own game, in this case perhaps refining your hypothesis to further account for your table image, your opponent's distraction and course your opponent's possible stupidity.

My opinion is that the true EV of a poker decision should be measured not only with the expected result but also the expected future value of the information you receive about and convey to your opponent and the value to you of learning about the efficacy of the play in all future hands, minus the expected future cost of any additional information you have given your opponents and, one hopes not but it happens, the expected future cost of any false information you "learned" about the efficacy of the play in all future hands. These added elements are very small I think compared to the instant EV and generally should not effect specific decisions except at the margin (Sidebar noting Bob Ciaffone's excellent point that advertising is a natural byproduct of correct play and should never be done for its own sake; I think this also applies to information gathering and learning), but they are real, and if you react by ignoring the potential negative evidence against your game, you are leaving money on the table. It is plus EV to learn (thus the title).

Finally, I agree with Fluxer completely that Sklansky's FTOP is a far from perfect guide to actually playing poker. However, I do believe that deviations from it are usefully analyzed by reference to it--i.e., when you make a move that you wouldn't make if the cards were face up to induce your opponent to do the same and your opponent makes a an FTOP correct call, you need to at least consider the possibility that you made a mistake. Losing the hand is an indicator that you should do that. In fact I think it is imperative that you do so, because I think we are wired to find the fault with the other guy and if we do that we are not going to get any better from our mistake, which itself is additional -EV.

But sometimes you still got to ask, how the hell could you make that call???


At Tue May 23, 07:21:00 PM 2006, Blogger DP said...

I typed a response here and then posted it on my blog, after realizing it was too long to be put in the comments.

At Wed May 24, 10:56:00 AM 2006, Blogger Raveen said...

see the point you make is an old point nothing new in your entry but what you dont realize is that what you think is a good play, and what i think is a good play are two different mind sets. Depending on your level of experience, the limits you play, and your ability to assess situations good plays are circumstantial. Im not saying that no one makes good plays but im saying its tough to define what a good play is. Is a good play raising 2X the BB with AA and getting it all in on the flop only to lose to a guy calling with a flush draw with no odds. Or is a good play limping with AA giving the guys odds on every to catch his flush, checking it to him when he missesd and having him go all in on a bluff which you call down. Beginners might look at the both situations as unlucky when in fact they werent' they were just straight up bad plays...those examples arent the best but there what i came up with on the spot but i hope you see my point


Post a Comment

Links to this post:

Create a Link

<< Home