Risking Your Whole Stack on the Flop

[EasilyFound]

In two spots, NLTAP gives examples of when it is "okay" to commit your whole stack, on the flop, to a hand. In the first, it is with a set if 5s with 100 bb in a limped pot. In the example, the board is K 9 5, with two hearts. In discussing how to get your opponent's final $480 into the pot, the book at pp. 28-29, says to ignore the possibility of your opponent having a set of 9s or Ks, stating that you're "destined" to lose your stack in such a case. The book says that this is a "destined" hand because of the size of the starting stacks, and that your destiny would be different if the stacks were ten times deeper ($5000).

In another hand, discussed at 41, the book gives an example of a hand in which the player is likely to be betting his entire stack before the end of the hand. The hand is AQ on a drawless board of Q 8 3. In the example, the player has $800 left in a pot of $100.

Sorry for the long set-up, but I guess these examples started me thinking about how deep you'd have to be in the second example before you'd not be "destined" to lose your stack.

Many of my cash home games are played with very shallow stacks, where players are likely to be shallow on the flop, so questions like this are important. I understand, of course, that sometimes you have a read that would change your decision, but I'm curious what the baseline would be for such a situation.

[cynic757]

I recommend that you check out Professional No Limit. It just came out last month and addresses the exact question you're asking.

[EasilyFound]

You're talking about the book by Flynn, Mehta, and Miller?

[cynic757]

That's the one. If you want to get a feel for it check out the threads currently going. One is in "Books and Publications" and I believe a post has a link to the others.

[mce86]

In the second example, you are not destined to lose your stack. What he is saying is, that in deciding your bets, if you plan on going to war with this hand (will call a big bet), you have to make your flop bet big enough so that in the long run, even if youre opponent draws out, he will be losing money by making the call. To do that you have to include your stack in the pot when determining what odds you are giving him.
The first example you are destined to lose because you arent going to fold a set in this spot. You could fold AQ though on that board to heavy betting
To sum it up, if you are going to pay off big bets, make your flop bets larger to take away some of the implied odds so that even if he hits and wins, its not a good call.

[EasilyFound]

[ QUOTE ]
You could fold AQ though on that board to heavy betting


[/ QUOTE ]

I did not read the example that way, since the book says, "With such a strong hand and so little left to bet, you likely will be betting your entire stack before the end of the hand." I see what you are saying in the rest of your post, but I took the example to mean that given the stack-to-pot ratio, the player should play the hand to get all his money into the hand. The book does not discuss how to play the hand if the opponent shows strength after the player moves strongly at the pot with intent to get all of his money into the hand before it ends. I guess I assumed, based on the quote, that the player would be pot committed.

[Mook]

[ QUOTE ]
I guess these examples started me thinking about how deep you'd have to be in the second example before you'd not be "destined" to lose your stack.

Many of my cash home games are played with very shallow stacks, where players are likely to be shallow on the flop, so questions like this are important. I understand, of course, that sometimes you have a read that would change your decision, but I'm curious what the baseline would be for such a situation.

[/ QUOTE ]
I haven't read PNL yet, so it may be that this question is answered in detail there. But I'll give it a shot regardless.

The "destined to lose your stack" comes from the math underlying set mining. I'd like to run the exact numbers using PokerStove, but I'm at work, so approximating it will have to do for now.

You'll flop a set with your PP approximately 12 times in 100. However, your opponent (if he holds a higher pocket pair) will simultaneously hit a bigger set about 1 of those 12 times. With 100 BB stacks, if we can make as little as ~10BB off our opponent on average those 11 out of 12 times he doesn't have a set, we could stack off our 100BB every time he did and still come out ahead in the long run. Which shouldn't be tough, especially if you follow the bet sizing techniques advocated in that section of NLHTAP. (And because the pot on the flop is already ~8BB, since IIRC the example Miller used was a raised pot with you holding 44 vs. Villain's AK, not a limped pot.)

But let's say the stacks are 1,000BB instead of 100. Now, if we were willing to wantonly go broke with bottom set, we would need to make about 90BB on average - more than ten times the current pot - from our opponent every time our bottom set is good. Which is next to impossible, given the number of times he'll give up instantly (e.g. if he holds 66 on a K94 flop) or maybe after a single pot-sized bet (e.g. if he has QQ on that same flop).

By plugging in a few assumptions about how tight and how aggressive your opponent is, you can come up with a number for the "threshold" past which pure set mining begins to break down from an EV standpoint. I'd guess it would come at somewhere around 300BB. I might try to play with the numbers a little bit later on.

Mook

[EasilyFound]

I see. My home games are never very deep, about 40-50 bb per buy-in, and I don't play away from home, at a casino, where games have stacks that are deeper, so if I hit a set in my home game or in a cheap tourney, I don't worry about set-over-set.

I'm thinking more about the common situation of where you have top pair, top kicker. But I suppose there are too many variables to have a similar rule for that situation, such as knowing the tendencies of your opponents, and the number of potential overpairs, and any flopped two-pairs, etc . . .