Thursday, September 27, 2007

Continious improvement

There's a concept in operations management called continuous improvement.

One approach is

Plan: Identify an opportunity and plan for change.

Do: Implement the change on a small scale.

Check: Use data to analyze the results of the change and determine whether it made a difference.

Act: If the change was successful, implement it on a wider scale and continuously assess your results. If the change did not work, begin the cycle again.

I prefer a slightly different approach for a continuous improvement plan in poker.

Plan Identify a potential problem area or opportunity for improvement.

Analyze Collect data and make observations to determine the extent of the problem or opportunity and potential improvement

Implement If the potential for improvement exists then make changes

Evaluate changes If the change doesn't result in improvement then redo the process to make sure you correctly identified the problem/opportunity.


Poor Danny

Danny Boy explains a hand he played in a recent tournament.
My last hand actually went like this: I had just raised two hands out of the last three and won both with no contest. Based on that, I limped in under the gun at an 8-handed table with A-Q, also because I knew the big blind call my 2.5 times the blind raise. The button made it 2500 and it felt to me that it was simply a position raise. THE SMALL BLIND CALLED and I called.

He limped because he knew the big blind would call?

He limped because he knew he'd get 2 to 1 on a raise?

Has he quit drinking or something? If so, he needs to start again.


Wednesday, September 12, 2007

Another tell on TV

F-train mentioned this verbal
tell by Hellmuth
on High Stakes Poker the other night. It's a good one. The boy just talks to much.


Sunday, September 09, 2007

The 80/20 rule

Wayne Vinson doesn't think much of my blog. In a comment on a previous post he says
Well, good for you to man up and admit you were wrong. Now you need to tackle about 80% of the rest of this site, because your errors are rampant, not localized.

I'm not so sure that 80% of what I say is wrong. But, even if it is, does that actually mean anything? I don't think it does.

The old 80/20 standard says that 80% of the value comes from 20% of the information. What that means is that the stuff I'm wrong about is likely stuff where being wrong doesn't cost much, where the difference between right and wrong is small. But the stuff I'm right about is stuff where it matters, and it's also not superficially obvious.

So it could well be that even if I'm wrong 80% of the time my thoughts have more value than someone else who's only wrong 20% of the time. It just depends on which 80% and which 20%.

It would be helpful if Wayne offered more information about which 80% I'm wrong about. Maybe he'll do that.


Pot odds -- a clarification

In a comment to a previous post Wayne Vinson says
Harrington uses different terminology than you do, Gary. Specifically, "pot odds" are a broader concept in the way he defines them.

He's making that comment as a result of my having said
On page 14 of Harrington on Hold 'em Expert Strategy for No Limit Tournaments, Vol. 1: Strategic Play they say

There are a variety of mistakes one can make in poker, but one of the most serious is to make a bet or call which is not correct given the pot odds available to you ...

That's just a fundamentally wrong statement that I'm not sure how a book that starts out with that idea has any credibility at all. The power of TV I guess.

The stunning error in the statement is about the idea of betting without the correct pot odds. Pot odds have nothing to do with betting. At least not your pot odds.

Wayne's comment caused me to go look up how Harrington does define pot odds. My comment was about something he'd said on p.14 without having defined pot odds so I just assumed he meant the term in it's generally understood meaning.

Harrington does define pot odds later, on page 122. He says
If your opponent has put you all-in, or has made a bet which is the last significant bet of the hand, then the pot odds are easy to calculate. Just calculate or make your best estimate of what's in the pot, and compare it to the amount required to call.
While it's true that Harrington uses more words in that definition than needed, it's not a broader definition than is the norm. If anything it's a more restrictive definition than normal, not a broader one.

Harrington goes on in that chapter to discuss potential changes in odds if there's a player behind you who might call or raise, and to discuss how pot odds might be modified to account for implied odds.

But there's nothing in his definition that suggests he's using the term pot odds in any abnormally broad way.

Maybe Wayne can clarify his comment?

It is true that Harrington tends to use language rather loosely. His phrase "when your opponent has put you all-in" is an example. Your opponent might have bet enough to put you all-in, but he hasn't put you all-in. He's bet enough so that if you call you're all-in. You still get to decide whether to call. That's a subtle looseness of language, but it's still an example of Harrington not getting it exactly right.

So, it's probably true that Harrington does use the term pot odds in ways outside the generally accepted meaning of the word, and outside his own definition of the term. But that just means he's sloppy with his thinking, which was my point in the first place.

Labels: ,

Anchoring bias and adjusting your strategy from a base

I've often argued that it's not a good idea to think in terms of a base strategy that you adjust to meet changing conditions. I'm not sure I've ever made a good arguement or explanation for why I think it's a mistake to do that though.

So I thought I'd give it another try.

Rational thought is not the norm for human beings. There tends to be natural bias of many different kinds that bends the way we make decisions. There's probably evolutionary reasons for this. But the reasons for it don't really matter. We do tend towards bias in the way we make decisions, in the way we think, and if you're a fan of the idea of rational thought it's important to recognize these natural biases and to guard against them.

There's a human decision making bias that psychologists call anchoring bias.

Let me explain it with an example.
Suppose I spin a Wheel of Fortune device as you watch, and it comes up pointing to 65. Then I ask: Do you think the percentage of African countries in the UN is above or below this number? What do you think is the percentage of African countries in the UN? Take a moment to consider these two questions yourself, if you like, and please don't Google.

Tversky and Kahneman (1974) recorded the estimates of subjects who saw the Wheel of Fortune showing various numbers. The median estimate of subjects who saw the wheel show 65 was 45%; the median estimate of subjects who saw 10 was 25%.

If I try to determine the best strategy for a given situation by a focus on adjustment from some baseline strategy you'll likely not arrive at the best strategy, your judgement will be biased no matter how hard you try to compensate for the bias.
Debiasing manipulations for anchoring have generally proved not very effective. I would suggest these two: First, if the initial guess sounds implausible, try to throw it away entirely and come up with a new estimate, rather than sliding from the anchor. But this in itself may not be sufficient - subjects instructed to avoid anchoring still seem to do so (Quattrone et. al. 1981). So second, even if you are trying the first method, try also to think of an anchor in the opposite direction - an anchor that is clearly too small or too large, instead of too large or too small - and dwell on it briefly.

The best approach for poker players is to focus on the current situation rather than a baseline strategy. What is it about the current situation that creates value or risk? Don't ask yourself How should I play suited connectors in general and how should I adjust that for my position? Ask yourself What is it about my current position that creates value then look at your hand and see if you have a hand that can exploit that value.

It really does make a difference how you think about something. It might seem silly and trite and unimportant to think about things in a certain way, but it's important if you want to avoid mistakes generated by psychological bias.


Flopped full house

Looking for something to blog about I ran across this poll on how to play a flopped full house.

The situation is that you're in late position with a T9 and a flop of TT9, two suited. There are 4 other active players, the UTG had limped preflop, two callers, you called, the BB checked.

Now, on the flop, the UTG bets 5 (it's a 1/2 game), there's two callers, the BB is still behind you. There's plenty of chips. One of the callers has $70, everybody else has a couple hundred, you have everybody covered.

What do you do? The poll gives three options, call, raise a small amount, raise a large amount.

70% say call
25% say small raise
5% say large raise

I think call might be right, but I don't think it's clearcut. It's the first impulse of most players, it's my first impulse.

Aggression in no limit games is often overrated. Passive play often gets more money than aggresive play. But at the same time slowplay is overrated also.

The more I think about it the more I think it's a good chance one of the 3 hands that have already put money in the pot this betting round has a T. If that's the case and they're slowplaying it then a small raise might well get them to come out of the woodwork. A small raise might well have a good shot at getting all the chips in right now against one player with at most 3 outs.

Raising might scare off the flush, straight, and overcard draws. But if somebody else has a T and you're both slowplaying it then I think that might well be as big a disaster as driving off the draws. You aren't likely to get all the chips of a flush draw even if he gets there.

Then again, against another T you can probably get all the chips in on a later street.

In a limit game I think a raise is clearly the right thing. In a no limit game I don't think things are ever that clear.


Saturday, September 08, 2007

Calling on the button

The UTG player makes it $10 in a NL 1/2 game and everybody folds to you, on the button and you have QhJh. What should you do?

Well, of course, it depends. It depends on a couple of things. What is the range of hands the UTG player would be making that raise with? How likely is it that the UTG player has a big pair that he'll fall in love with? How deep is the money?

The deeper the money, and the more likely he is to fall in love with his hand, the more likely you should be to call. It's harder to quantify that than the effects of the range of hands he might have though. So, let's just look at what's easy to look at, his range of hands, and use Poker Stove to try to map out a general approach.

1. If he's a very tight player and you can put him on QQ+ or AKs then he's pretty far ahead. He has over 78% equity against a QJs.

In this case you're going so far uphill that even with very deep money and an opponent who's very likely to fall in love with his hand it's probably not worth trying to climb that hill. Just give it up.

2. But most UTG openers won't have that tight an opening range. Even most fairly tight players might have AJs, or AQo, or even 99. Looking at that range his equity against QJs is only about 69%. He's still pretty far ahead, but not so much that you might want to take a shot if the money is deep and he's the type to fall in love with a hand. The money does need to be pretty deep though. Probably an extra $200, which isn't always available in todays capped buyin 1/2 games. And just deep money isn't enough, he also needs to be a stubborn type who won't give up the hand when he's clearly beat.

3. How about a more typical player? One who will come in with KQo and pairs as low as about 77 and even suited connectors down to JTs? How does your QJs look against that range? In that case his equity has dropped to about 63% and you're not in such bad shape. Now the money doesn't need to be as deep, maybe $100 left in the stacks is enough, and he doesn't have to be 100% stubborn, just a good chance he'll fall in love with his hands is probably good enough, you don't need to be sure of it.

4. There are players who are even looser. They'll come in UTG with that raise with hands like KJo, or even 76s or KTs. Against a player like that you're in pretty good shape, he's about a 60/40 favorite. Stack sizes as small as $60 or so is probably enough.

The main thing to learn from this short analysis is that you aren't ahead with QJs against an UTG raise, even with your positional advantage you're going to be starting out behind. You're going to need fairly strong implied odds to justify a call even against some of the loosest early raises. Most of the time you're probably better off just passing.

Labels: ,

Get some rest

Random Shuffle has a post about not playing when you don't feel up to it.

I support that line of thinking completely. I think the single most important controllable element for a winning poker strategy is to play only when you're feeling your best.

And I don't mean feeling good because you've been doing to much cocaine.