Shortstacked 99 in big blind facing raise

[bucktotal]

[ QUOTE ]

Now, there could be times when you want to reduce your variance. For example, you're at the FT, there are a couple of short stacks, and the money increase is significant. You'd be willing to trade a little cEV for some $EV. This isn't one of those times.

[/ QUOTE ]

push PF

[allenciox]

Ok, let's look at what happens if you stop-n-go in each of these:

1) You hit set (12%). Let's assume you milk it like you would in the case I mentioned earlier, rather than automatically going all-in here, which would be really stupid. You have an expected gain of 1701 TC.

2) The flop comes A x x (8% of flops): Let's assume that they fold everything but a hand containing an Ace or KK or QQ. 35% of the time we win 900. 65% of the time we go to showdown, of which we lose 1952 90% of the time and win 2452 10% of the time. In this case we have a net expected result of -668 TC.

3. (29.6% of flops). Assuming opponent calls with a piece of the flop or with a number of outs, he will fold only about 20% of the time here. When he calls, you have about 20% equity (using Pokertracker with a number of combinations). So your net result is .2*900 - (.8)(.8)1952 + (.8)(.2)2452 = -678 TC.

4. (16% of flops) Same result as my suggested approach, net result -639

5. (34.4% of flops) Assume that they call your all-in with AK and all pairs (49%), and fold everything else. So 51% of the time you win 900, and 49% you go to showdown, 40% of which you win 1952 (by Pokerstove) and 60% you lose 2452. Your net result is -231.

So your net result with the stop-n-go is .12(1701) - .08(668) - .296(678) - .16(639) -.344(-231) = -232 TC.

So the stop-n-go is not only worse than a raise all-in (14 TC) but much worse than a fold in this situation!

[woodguy]

[ QUOTE ]
I think its safe to say that it's close. By the Nath principle, we should do whatever is more aggressive in that case.

[/ QUOTE ]

This sounds like Jason Strasser's axiom of when confused, its probably best to push.

Regards,
Woodguy