Mason's article on bunching

Was excellent. It made me think of something new.

Generally, bunching is described as a way to predict the remaining cards of players yet to act, given that players ahead have folded. Mason has shown this effect is negligible to unmeasurable.

OTOH, bunching--or perhaps more accurately--"reverse bunching" can be used as a way to deduce what the players ahead of you do NOT hold. (Mason's example of over cold-calling a preflop raise with a small pair in position.) I never thought about this aspect, and it was helpful.

[/kissass mode]

[mojobluesman]

He's right about one thing. Other people have thought about this conceptually. I know I have and I'm not even all that good a player yet. However, there's a big difference between identifying "issues" like these and translating them into the correct responses at the poker table. This is the kind of article that is helpful to a "thinker" that hasn't put it all togther yet.

[belloc]

It's often brought up in a game like O/8, where there are more cards dealt, and only a handful of "good" cards. So, when you're in late position in an otherwise typical game, and six players limp in front of you, you can be pretty sure there are a lot of aces and deuces in those hands, which might help you make a better decision when it's your turn to act.

I've never seen bunching discussed effectively in holdem, but Mason's examples seem to be the right sorts of situations where bunching might help you decide one way or the other.

[ZeroPointMachine]

Another common O8 example that I see late in tournaments regularly. A hand is folded around to a late position player who makes a stand/steal with some KKxx hand. The aces turn up in the blinds way more often than expected.

[dfan]

Mason, while the concept is a useful one and I appreciate you pointing it out to me (I had never thought of it that way) your guestimate as to what it does for your odds is off, at least in your first example. Also (unless I'm missing something) the math is simple and I'm surprised that you didn't recognize this.

The situation was a raise and 3 callers in front of you. You argued that none of these 4 players is likely to have a 3 given that they entered the hand.

Lets assume the best case, that these 4 players never have a 3. That means that instead of their being 50 unknown cards 2 of which are 3's (which gives you the usual 7.5 to 1 odds), in this case there are 42 unknown cards containing 2 3's. Then the prob of hitting a set on the flop is:
1 - (40/42 * 39/41 * 38/40) = 13.9% or odds of 6.1 to 1. While the bunching effect does improve your odds, it does not reach the 5 to 1 number you guestimate.

Also the 6.1 to 1 assumes that those 8 cards in your opponents hands NEVER includes a 3. Since this won't be true your true odds of hitting your set will be slightly worse.

[dfan]

Never mind - I see this has already been pointed out in another thread.

[Beavis68]

I didn't like his "reverse bunching" example. 10:1 to hit a set means that there would have to be 33s 50% more often in the discarded hands that typical. Statistically there should be about .7 3s amoung the folded 18 cards, he is predicting a little over one 3, in the 18 cards.

[winky51]

Good article. I already was using this strategy myself. I am always looking for loop holes. But his 5:1 is a bit off. Id say at best its 6:1 or 6.5:1 but thats still good. Reverse bunching is possible but players will fold many hands. But all it takes is 1 three to be tossed out to really shift your odds of a set. 1:50 instead of 2:50 with 3 cards to come. And yes the other player will miss 2/3 of the time he is stealing your blind but what are you going to call or not call? AJ9 with 2 hearts flop, now what with a pair of 33s. Or 874 flops, now what?

Its touch defending with that hand in limit.

[mojobluesman]

I think the 10-1 referred to what you need to make the hand playable because you will hit your set and lose sometimes.

[me454555]

I have to disagree w/masons 33 blind steal situation.

In his blind steal situation, I set it up to give button the top 30% of hands and to give you 33. Your equity is 46%

equity (%) win (%) tie (%)
Hand 1: 46.3419 % 45.92% 00.43% { 33 }
Hand 2: 53.6581 % 53.23% 00.43% { 55+, A2s+, K5s+, Q7s+, J8s+, T8s+, 98s, A7o+, A5o, K9o+, Q9o+, J9o+, T9o }

I set it up to run a second time and this time I made the 3c a dead card to simulate his bunching effect. You still have 41% equity

Board:
Dead: 3c

equity (%) win (%) tie (%)
Hand 1: 41.0308 % 40.59% 00.44% { 33 }
Hand 2: 58.9692 % 58.53% 00.44% { 55+, A2s+, K5s+, Q7s+, J8s+, T8s+, 98s, A7o+, A5o, K9o+, Q9o+, J9o+, T9o }

You're getting 3:1 on your initial call and if we assume a rake of 1 sb, if we took it to showdown we'd be getting 4:3 so our equity would be 42%. Given that 1) the 3c will sometimes still be live for us and 2) rake will not always be 1 sb in different games and 3) Our opponents can be more aggressvie than a 30% attempted to steal blinds, I think this could be a fine call in the bb.

Also note that we can eveluate our position after the flop and that can also improve our equity since we will not go to showdown if the flop is AKx or so.

Calling 33 a fold in the BB is not necessarily correct and I wish it would be expanded on more.