I understand this. Is there any reason why this should be at least double your winrate?
Take the number PT reports in the BB/Hand column with the "On Steal" filter applied and multiply that number by 100 to get your BB/100 on steals. Compare this to your regular BB/100 winrate on all hands.
Meh, double was arbitrary. If I were writing it now, I'd say it should be more like quadruple, but that's still just a vague rule.
If I really wanted a number to compare, I'd probably take the reported per-hand winrate, multiply by 100, and multiply by my VPIP (as a fraction). I'll use my numbers as an example:
Over the last 3,710 hands at $100NL, I've had a winrate on blind steals of 1.65 PTBB/hand.
Multiplying by 100, that's a winrate of 165 PTBB/100.
My VPIP has been 26.77%, so multiplying we get 165*(0.2677) = 44.2 PTBB/100. That should be larger than my regular winrate.
The reasoning behind this math is that our winrate is over the hands that we actually *play*. If we have a winrate of 6 PTBB/100 but a VPIP of 20%, that means that on the 1/5th of hands we actually choose to play, we win an average of 30 PTBB/100, and then on the 80% we don't play, we win nothing. Since the "blind steal" filter automatically focuses exclusively on hands where we are playing, our winrate should be much higher -- multiplying by your VPIP gives you more of an apples-to-apples comparison.
The main point is that you need to compare your blind-stealing winrate to a higher number than your overall PTBB/100 or else you're deceiving yourself into thinking you're doing a great job when you might not be.
Blind steals should be VERY profitable since you're in the most advantageous situation that the game can offer. If you can't post a SOLID win on your blind steals, you need to reconsider your strategy.