5 ms·
It seems like a significant confounding factor in analysis like this could be the "natural floor" of a beginning player. Some people signing up for Lichess have
by georgeoliver 5y ago
It seems like a significant confounding factor in analysis like this could be the "natural floor" of a beginning player. Some people signing up for Lichess have never or barely played a game, others have lots more experience. Wouldn't this affect their rate of improvement but be invisible in the data set? The OP touches on this a bit; what would be good ways to filter for it?
- LorenPechtel 5y agoYes, I think the data is inherently corrupted this way. There's nothing to indicate when someone started playing. I'm on Gameknot rather than Lichess, but you try to measure my improvement? You're not seeing the 30 years I was playing before I ever did it online.