Today I wrote a utility to run my new indicator on my existing poker session data. I decided that the only goal of the indicator should be to identify winning sessions, not to classify each session as either winning or losing. In other words, I consider it a worse failure of the indicator if it gives a false positive than if it fails to identify a winning session. Here are the preliminary numbers:
471 number of no limit Hold'em sessions for which I have the full hand histories
329 number of these sessions which were winning ones
117 number of sessions which the indicator flagged as winning
24 number of false positives flagged
79.49% accuracy rate of the indicator
As I stated in yesterday's post, the indicator should only be run on sessions in which I did not hit the felt. I need to modify the utility to exclude those sessions where I did hit the felt. My guess is that once I do, the corrected accuracy rate of the indicator will be higher than 79.49%.
During current Hold'em session you were dealt 45 hands and saw flop:
- 4 out of 5 times while in big blind (80%)
- 6 out of 6 times while in small blind (100%)
- 20 out of 34 times in other positions (58%)
- a total of 30 out of 45 (66%)
Pots won at showdown - 4 of 5 (80%)
Pots won without showdown - 2
delta: $16,819
cash game no limit hold'em balance: $4,880,896
balance: $7,237,849
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment