re: Poker & My 10NL experience (August, 54k hands) - systemic mistakes or just "variance"?
Honestly, your point is valid if there isnt a sample size concern, but the sample is not 54k hands because you're trying to compare the 18 times you've had set over sets or something.
Also, you are really really really over estimating the amount of bb/100 you're losing by keeping this leak, if it is a leak then its a small leak at best due to how rare the situations are, nowhere near the magnitude of even 2bb/100, if it EVEN is a leak, i've done quite a lot of database work myself but i still can't even check the profability of myself playing small pairs with any decent reliability, and i've got over 300k hands at zoom, and that's just overall profitability. You're narrowing down the analysis from the profitability of small pairs, to set over set spots with small pairs, and if you can't already see what im getting at, its super meaningless - 18 times in 54k hands, id guess its just as likely that you'd create a leak by stopping set mining at these limits, to be honest.
Im not saying there isnt times to drop small pairs, infact theres tons of them, so work on folding them when there are too many players to act behind, there is a big squeezer behind.. etc, but folding them for the most part especially in late position at nl10, or vs nits, is so much more likely to create a leak than calling with them, as they tend to be very profitable hands at the micros for most people.