1) Some granularity into the density functions of the rating distribution by format. I know on https://www.chess.com/leaderboard/live/bullet one can get approximates and probably manually create the existing distribution. But it’d be nice if someone already has an excel sheet or Shiny app or something
2) How these distributions have changed over time. For example, I know a few years back, they gave everyone a boost in rapid time controls since the relative distributions were so lopsided. But again looking for more of a granular look into the data, e.g. a month-to-month look at the given CDFs complete with active member data.
This is spurred on by three things: seemingly much easier matchups at the ~2000 rapid level (am I just getting lucky/heavily improving there, or has something changed), an uneven distribution of the daily ratings (>40k people are above 2000 in blitz and rapid, but only 3k are above that mark in daily), and a general interest in statistics.
I am looking for two things:
1) Some granularity into the density functions of the rating distribution by format. I know on https://www.chess.com/leaderboard/live/bullet one can get approximates and probably manually create the existing distribution. But it’d be nice if someone already has an excel sheet or Shiny app or something
2) How these distributions have changed over time. For example, I know a few years back, they gave everyone a boost in rapid time controls since the relative distributions were so lopsided. But again looking for more of a granular look into the data, e.g. a month-to-month look at the given CDFs complete with active member data.
This is spurred on by three things: seemingly much easier matchups at the ~2000 rapid level (am I just getting lucky/heavily improving there, or has something changed), an uneven distribution of the daily ratings (>40k people are above 2000 in blitz and rapid, but only 3k are above that mark in daily), and a general interest in statistics.