I have a formula that I use where I combine expected value and baysean probabilities. It's intense and it requires focus, risk and adversarial thinking, but it's working out pretty well for me.