In case anyone is interested, here are the (rating, percentile) pairs I collected for this plot:
(514, 37), (536, 39.4), (545, 40.4), (570, 43.1), (579, 44.1), (581, 44.3), (585, 44.8), (589, 45.2), (600, 46.4), (613, 47.8), (632, 49.9), (638, 50.5), (661, 52.8), (663, 53.1), (668, 53.6), (685, 55.3), (699, 56.7), (711, 57.9), (718, 58.6), (727, 59.5), (738, 60.6), (759, 62.6), (790, 65.5), (819, 68.1), (833, 69.3), (866, 72), (918, 75.8), (967, 79), (1099, 86.3), (1192, 90.1), (1447, 96.4)
That's interesting. But on some accounts I am below 1000 and can't beat 800s easily. On other accounts I am ~1400.
I don't think you are collecting the right data. For example, when did you collect this information? Did you compare the performance of players at 3am equally as those playing at 3pm?
If you are in America, you most likely got a slice of American stats, not European.
Another factor is tournaments. I noticed players are not as strong in tournaments as they are with random pairings.
Then you have the one trick ponies. They are good in one line but they don't do so well in the rest, which is keeping them at a low rating. This skews the results for other players at the same low rating level.
Game accuracy would be a better measure than rating in my opinion.
Hey everyone,
I thought it was weird that there isn't a more direct way to view information about the rating distribution, so I gathered some manually and thought it could be of interest to the community. I just checked the stats pages of other players across the rating spectrum and recorded their ratings and percentiles, and then plotted the results and drew a curve. I'm sure there are mistakes, but it's a useful way to see at-a-glance how players are distributed across rating bands. Hope this helps if anyone else is wondering about the same thing!