Correction on the last post. You can also measure correlation without looking at a specific individual over time, the average APM and winrate would be for everyone in the total sample, not for this specific individual. The calcuations will be the same though, and it won’t change my general point.