[ QUOTE ]
[ QUOTE ]
Calling is better than folding if your expected return is greater than the size of your call. Your expected return is the sum of the expected return from two cases.
[img]/images/graemlins/spade.gif[/img] Case 1 is that at least one card of your rank is on the flop. This is fun, although you only win about 80% of the time, given that your opponent started with an overpair.
[img]/images/graemlins/spade.gif[/img] Case 2 is when you miss. This is usually uncomfortable, but you might miss with a straight-flush draw and 17 outs against an overpair without a card of that suit.
[img]/images/graemlins/spade.gif[/img] Case 3 is that you don't get to see a flop, perhaps because there is a big reraise after you. Your return should be 0 for this case, but it can't be ignored because its probability affects the probabilities of Cases 1 and 2.
[/ QUOTE ]
pz, great post.
[/ QUOTE ]
Thanks. This approach is much more common in backgammon analysis than in poker. In backgammon, you may need to find 25% to take a double (a 1/2 pot bet), and you might get hit immediately 14 times out of 36. A common technique is to count your wins out of the 14 hits (gammon losses are -1/2), and your wins out of the 22 misses. You try to come up with the 9 wins out of 36 you need to take. Usually, you have to consider both cases. I think this method is effective for reducing the complexity of evaluations, and could be applied to poker more.
For example, if you are thinking of completing the small blind with a hand like T4o in a family pot, there are 3 cases again. Case 1 is that the flop brings 2 or 3 cards of your ranks, which happens about 3.5% of the time. Case 2 is that you hit the board at most once (including the times you have a 1 card draw). Case 3 is that the big blind raises, and you never see the flop. To expect to get your completion back, you need to find enough value from a combination of Case 1 and Case 2, and usually you can't.
Here is
another example.
In poker, it is common for people to assume implicitly that they are dead in Case 2. If you have an OESD on the flop, you don't know that your opponent will fire again on a blank turn, or that he will bet enough that you will not have at least a profitable call.
[ QUOTE ]
are you assuming that your opponent in Case 1 & 2 is a good or at least equal opponent? I ask because often we talk about finding opponents who make the most mistakes while playing, and exploiting those mistakes. I think that Case 1 & 2 change depending on your ability to exploit bad player's mistakes, or whether you're up against a smart player.
[/ QUOTE ]
You can and should combine this with assessments of your opponent's mistakes, and how well you believe you can exploit them. An opponent who will stack off too often will increase the value of Case 1, while decreasing the value of Case 2. A looser opponent may pay off less in Case 1, while giving you more in Case 2.
While you may want to tailor your evaluations to your opponent, it would be nice to see overall statistics. I don't know how to extract these from PokerTracker, but it would be a useful feature.