Bluffing into dry sidepots...

[Red_Diamond]

God knows I've had my own problems with donkey plays this week. I had a couple situations where some euro-donk bluffed into a dry sidepot (even on the river!), and let shorty quadruple up which came back to eliminate the rest of us.

While I can't understand such foolishness in tournament poker, I have read over the two concepts by Chris Ferguson regarding this case. And surprisingly, there is evidence to support that bluffing into dry sidepots can be +ev. This is due to the fact that a third player must call with even weak holdings to make sure shorty doesn't live. This in turn, allows the bluffer to make profit when he DOES have a hand, as it is never known for sure.

These papers always worry me, as I am afraid they may cause more of a trend of this behaviour. I can only assume Ferguson was trying to disprove the bluffing aspect, but instead ended up proving the opposite to his intent?

For those who have read Bill Chen's work, I'm curious what you think.

[AaronBrown]

It is often correct in tournaments to bet into a dry sidepot. It rarely makes sense to run a complete bluff, but you don't always need a really strong hand to go all-in.

Let's take a simple example. There are three players left in the tournament with remaining prizes of $10,000 for first, $7,000 for second and $4,000 for third. Players A and B had 40,000 chips left at the start of the hand, player C had 10,000. Everyone is in for 10,000.

If A and B check and C wins the hand, all three players will be left with 30,000 in chips and $7,000 in expected tournament winnings. If A and B check and one of them wins, the winner will have 60,000 chips and $9,000 in expected tournament winnings while the loser will have 30,000 in chips and $8,000 in expected tournament winnings.

Suppose instead A goes all-in. If B calls and A wins the pot, she gets $10,000 instead of $9,000. If B calls and beats both A and C, A gets $7,000, the same EV as if she checked. If B calls and beats A but not C, A gets $4,000 instead of $7,000. Therefore, A gains from betting if B calls, as long as her chance of beating B and C is more than three times her chance of losing to B who then loses to C.

If B folds, and A beats C, she now gets $9,000. This is the best she could have done if she had checked in this circumstance, and it makes her an extra $1,000 if B's hand would have beaten hers. If B folds and A loses to C, she has $7,000. This is the same as she would have hand if C's hand were also better than B's, but it costs her $1,000 if C's hand were worse than B's. So if B folds, A profits from betting if she has better than a 50/50 chance of beating C, given that B beats them both.

Thus the key to betting in either case hinges on A's confidence of beating C. That's counterintuitive, because the bet is only with B, C's hand seems to be irrelevant. But if A has better than an even chance to beat C, a bet is often to her advantage. She should not run a pure bluff, but she only has to be a little better than C to act as if she has the nuts.

[Red_Diamond]

Aaron, thanks for the reply. I'm going to have to re-read your example over a few times and think on it some more. I see your case involves players that are already ITM, and not riding the bubble. But then again, they are riding other bubble-layers here nevertheless.

[rakemeplz]

I've thought about bluffing in sidepots is that even if it is not +ev to "bluff" sidepots you should still do it...this would be to get your value bets paid off by your deepstacked opponent who would be afrraid of transfering his equity to you or the shortstack. So even in some cases where its not EV it wuld be +EV across your range of hands. I'm sure (ok not really its all speculation) that sitations on the river would resemble what I've described above.

Anyways, could be wrong, might be right blah. Haven't thouroughly read the whole thread or Fergusons article's so these are my unfinished thoughts.

[AaronBrown]

Restricting it to three players just makes the math easier. The general principle applies. The farther you are away from the end of the tournament, the less the advantage to eliminating a player, and the more you should play for maximum chip EV. Also, the more confident you are about beating the short stack, the more likely you are to continue betting into the side pot.

Take it to the extreme. If you're certain of beating the short stack, then you have no reason to keep the other large stack in the pot in order to be sure of eliminating the short stack. You play just like a normal hand.

If you're pretty sure of beating the short stack, then the advantage of raising can outweigh the risk of letting the short stack survive.

The real reason to check when a short stack is all-in is it always helps the short stack for you to bet. If you were partners with the other large stack, you would never bet. The bet might help you, but if so it hurts the other large stack more. Refusing to bet in this circumstance so that others will not bet into you is positive EV, but technically collusion. It's so common that no one complains, and some people argue that it's part of tournament play.

[Recon20k]

i hate ppl who bet draws into dry side pots.. that wrecks my head. But if someone is gonna do it, whatever, either re-raise pre or check the nuts into them

[PantsOnFire]

Here is an example I have.

You are the chip leader with 10-12 players remaining. Only the final table of 9 will get paid so you are approaching the bubble. It may be +EV as the big stack to keep shortie in the game and continue stealing blinds and bullying to build your stack before this table (with people playing tight bubble poker) is merged with the other for the final table.