You also need to incorporate the "choose function", i.e. the number of different orderings of the i wins and 24-i losses. Then you end up with the binomial formula. This would give an approximation to the probability, but since the probability of winning changes on each hand, it wouldn't be exact. Also, you have to be pretty careful with these calculations for many reasons. For example, they weren't all heads up, it was probably more than 24 hands, etc. Especially if it "feels like 22/24". Also, it's very, very, very, unlikely that anonymous averaged an 80% favourite in those 24 hands. Very, very, very unlikely.