Tuesday, June 28, 2005

Update on my PokerTracker stats

Back in May I started using PokerTracker to track my results. I also thought it was important to understand how PokerTracker works since so many people use it and quote stats from it. (I will write more on the problem of relying too much on opponents stats when the sample size is small in the future) I posted some results in May based on 1000 hands at four different limits. Actually the last two limits are the same, one is 6-max, the other is full games. Here were the results (BB/100: Big bets won per 100 hands):

A. -7.09
B. +5.85
C. +11.67
D. +5.85

Net +3.56

At this point, I have played 17,000 hands, and now the updated BB/100 hands is:

A. +2.11 (4070 hands)
B. +1.74 (3185 hands)
C. +3.83 (10,092 hands)

Net +3.04 (17,347 hands)

I considated C & D into just C.

With more than 4 times as many hands played overall, its obvious that the results would be less wild and regress towards the mean. I don't think 17K hands is enough to show the true mean yet, but it should be in the ballpark. So the +11.67 and the -7.09 are much more realistic numbers now (+3.83 and +2.11 respectively), however, its possible the effects of those wild initial swings are still seen in the difference in numbers between A & C. A is a higher limit than C, so maybe the reason is players are better in A and the BB/100 should be lower. Maybe its a combination of those two reasons, very difficult to tell.

My comfort range in estimating my true BB/100 hands given the results thus far is a minimum of +2.00 and a maximum of +4.00. This assumes that I play the same ratio of hands in the three different limits. I would assume if I only played in the highest limit, that the BB/100 would start to tilt down to the lower part of that range. I wouldn't be surprised if it was lower than that.

It is really difficult to tell how many hands is enough to be a true gauge on one's abilities. With the information I have right now, I can only make some educated guesses.

No comments: