In long term, over a large number of issues, the outcome will converge toward the probabilities (as much as the model you're using is correct). Obviously in short term, and on any specific issue, you could have any outcome. So you have to correctly forecast the outcome of your current options positions. Also, the market prices in its current opinion, but this opinion continuously changes. The useful conclusion of all these is that it's futile to try to uncover an option strategy that would give you a positive expectancy, a strategy that would allow you to universally reduce risk and / or increase reward so you could apply it mechanically with success. Once you accept this conclusion you stop looking for an unreachable goal and you can concentrate on tangible targets.