What Is the Variance of Poker ?

Started by mpower 3 replies 1,980 views Jump to latest

mpower
mpower
  • 12 posts
  • Threads: 5
Joined: May 30, 2011
May 30th, 2011 at 4:30:52 PM #1
In Blackjack Appendix 4 we can get valuable information on standard deviation for the 6 deck blackjack shoe game. Did you make the same analysis for poker tournaments as well ?

My own research brought the following numbers:

1 Table Tournament with 9 players
EV = 0.1 Buy Ins (assumption for good player)
SDev= 1.7 Buy Ins

3 Table Tournament with 27 players
EV = 0.1 Buy Ins (assumption for good player)
SDev= 2.6 Buy Ins

Can you agree on the above numbers ? And what would be the standard deviation for a WSOP tournament with 1000 entrants ? Or even 8000 ?

mpower
PapaChubby
PapaChubby
  • 496 posts
  • Threads: 11
Joined: Mar 29, 2010
May 30th, 2011 at 7:31:54 PM #2
I'm using MGM Grand payout structure as a baseline. For both 9 and 27 players, the payout is 50% for 1st place, 30% for 2nd, and 20% for third. The buy in is $80, and 75% of the buy in money goes to the prize pool.

So the expected value is -0.25 buy ins in either case.

For the nine player tourney, the variance is 6/9*(-80)^2 + 1/9*(28)^2 + 1/9*(82)^2 + 1/9*(190)^2 = 9112. The standard deviation is the square root of the variance = $95, or 1.2 buy ins.

For the 27 player tourney, S2 = 24/27*(-80)^2 + 1/27*(244)^2 + 1/27*(406)^2 + 1/27*(730)^2 = 34202. Standard deviation = $185 = 2.3 buy ins.
mpower
mpower
  • 12 posts
  • Threads: 5
Joined: May 30, 2011
June 1st, 2011 at 4:39:53 AM #3
Thanks PapaChubby ! Seems pretty easy to adapt the variance calculation for any size of poker tournaments.

My intention is to define kind of a luck factor for poker tournament players. So far, the ratio between net win and standard deviation serves quit well.
PapaChubby
PapaChubby
  • 496 posts
  • Threads: 11
Joined: Mar 29, 2010
June 1st, 2011 at 6:53:51 AM #4
Yep, given the prize structure it's pretty straightforward to calculate the standard deviation. For a large tournament you'll probably want to use a spreadsheet.

I frequently use mean and standard deviation to determine the likelihood of being ahead or behind after a certain number of trials of a casino game. I use large bets and short gambling sessions to optimize standard deviation relative to expected loss. Based on the Central Limit Theorem, I assume a normal distribution for my analyses. You'll need to consider that the CLT does not apply to your poker problem, and the distribution is definitely not normal.

Join the discussion

Log in or create a free account to reply to this thread.

  • Jump to: