I have a set of data produced from a friend and I playing a game, where we take geometric trials of different probabilities, and we are trying to find an interesting way to measure "unluckiness" between the two of us. Example data: Pairs of coordinates are the % chance of success, and the actual number of trials that occurred. Friend 1: (17%, 33), (10%, 1), (7%, 10), (5%, 23) Friend 2: (17%, 12), (10%, 8), (7%, 7), (5%, 16) Is there a metric we can compute which measures who is "more unlucky"? We tried to compute the (1-p)^x for the probability of having at least as many failures, but how can we incorporate the fact the probabilities are different? This is purely for fun so any and all ideas are welcome!

Full article content could not be extracted automatically. Read the original below.