Identifying Leaks

[John White]

I'm looking at my pokertracker hands, sorted by reverse BB/Hand. I know that the big numbers (-2.4 BB/hand for K9s) are things I need to work on, but what's the threshold for being a problem?

Oh no. Do I have to calculate the significance of my sample size with that hand?

[sweetjazz]

Individual hands are statistically insignificant unless you've played several hundred thousand hands already. It's better to try to group them in a meaningful way.

If K9s looks like a prolbem, look at hands like QTs, K8s, KTs, Q9s, JTs, QJs, etc. Are you losing money with all of them? (If K8s is doing much better than K9s, what does that tell you about the sample size?) Look at individual hands and see if you played them right.

Here's a sample K9s suited hand (made up).

5 limpers to hero in the SB with Kc9c; hero raises; BB and 5 limpers call. Flop is Ac, 5c, 3h. Hero bets; 1 caller, EP raises; 2 more callers; hero 3-bets, 1 caller folds, EP caps, 2 callers call 2 more cold; hero calls. Turn is 8d. Hero check, EP bets, 2 call, hero calls. River is 8s; hero checks, EP bets, 1 caller, and hero folds. EP show As 3d, caller shows 6h 6c.

Hero played this made-up hand quite well in my opinion, and lost 4 BBs.

Or same hand, but the turn is the 2c. Hero bets, EP calls, MP calls; LP raises, hero 3-bets, EP, MP, LP all call. River is 3s. Hero bets, EP raises, MP folds, LP calls, Hero calls.

EP's A3 drags the pot. LP shows Tc, 4h.

Hero lost 8 BBs and played this pot well too.