You could try diving the number of wins by the amount the athlete is paid to find the best value athlete. Not sure it will give the answer that you are looking for because you're trying to reach two different and possibly exclusive goals simultaneously (poorest and most successful), but it would be interesting to see the result.
Unfortunately you'll have to examine your data set to find the best algorithm. For example, is payment correlated with success? Are they completely unrelated? How noisy and variable is the data?
Unfortunately you'll have to examine your data set to find the best algorithm. For example, is payment correlated with success? Are they completely unrelated? How noisy and variable is the data?
