Starting Bankroll for Kill Games

[jason1990]

I am going to also post this in the Poker Theory forum. For those who do not know what a kill game is, let me explain by example. I have played in a $2/$4 kill game in my local B&M casino. It plays like a standard $2/$4 limit holdem game with one exception. If a player wins two pots in a row, then the game becomes a $4/$8 game until a different player wins the pot. During this time, the blinds remain $1/$2, but the player who has won the consecutive pots must post a blind of $4. He will post this blind every hand until someone else wins the pot, at which point the game reverts to a normal $2/$4 game.

Using the recommended starting bankroll of 300 BBs, I should have $1200 to play in an ordinary $2/$4 game and $2400 to play in an ordinary $4/$8 game. So my starting bankroll to play in this kill game should be somewhere between the two. How much should it be?

As a related question, what is the long term ratio of kill pots to all pots? Or put another way, if I walk up to a 10-handed kill game at some random time, what is the probability that I will find a kill pot in progress?

And finally, how much should I buy in for when I sit down at a kill game?

[MentalCombat]

the % chance of a kill pot is 100/ number of players at the table.

I'd take that result, and multiply your bankroll by 1.XX

[Acepimp05]

Well the 300bb is not for one session its for a an overall bankroll online so it ensures you dont go broke, you should never sit down with that many chips at a little 2/4 kill game no one will wanna play a hand with you.People buy in at 2/4 at the mgm grand in vegas where i play 5 times a week for 200 and thats 50 big bets thats the right amount if your not a loose player who plays alot of hands to a showdown then i would recomend 400 to ensure not going broke.But dont sit down with 1200 at a b&m casino if ur good u dont want to be the hot shot at the table so u can keep taking there money all night lol. Where do you play at by the way is it in vegas or california or another state ?

[jason1990]

The kill games I'm talking about are in Milwaukee. Thanks for the feedback, but just so it's clear, I was asking two separate questions. One is what bankroll should one have and the other is how much should one buy in for. I usually buy in for 25BB in a non-kill game, so I figured I would buy in for 50 in the kill game, though that may be more than necessary. As for the overall bankroll question, that's separate. Obviously, I wouldn't sit at a table with $1200.

[Robk]

well taking the second question first. splitting it into two cases, raised pots and unraised pots. ill assume players raise about 9% of the time on average. (all of these assumptions are going to be ballpark figures btw.) that means the chances a pot is unraised are .91^10, or roughly 40% of the time. assume unraised pots are on average 6 handed, and the person posting the kill blind will have a slightly worse than average hand. so say he wins a bit less than 1/6 of the time, call it .15.

now for the raised pots. i ran a quick pokerstove sim for this one. i just set up a four handed pot and made some basic assumptions about what hands hero would play, the two other players, and the raiser, etc. the kill poster was winning the pot 23% of the time on average. again im not claiming this analysis is logically correct or even close really. as a close analysis of this would be complicated and probably not even get you that far from 23% anyway. lets call it 20%.

that makes the chances of a kill pot = .4*.15+.7*.65*.2 = roughly 14%.

of course the charcteristics of any particular game could change this figure a great deal. for instance, i have limited experience playing live poker but ive noticed that some players play much more tightly in kill pots. also the winner of the previous pot has a certain psychological advantage. all in all perhaps 15% would be a decent estimate.

then for the bankroll question just use the standard formulas. since you said 300BB was needed for the 2/4 player it looks like hes roughly a 1.5 BB/hr winner with 14 BB/ hr standard deviation (1% ror). just adjust these figures based on the estimate of how likely the pot will be killed (or whatever ones you had in mind) and plug them in. of course you could quibble about this and say that a good player wins fewer than his fair share of pots so hes less likely to post the blind which increases his winrate or probably a half dozen other things i havent thought of.