Rating Percentile Distribution Graph History

I hit 1493 for Daily at Christmas and at that time I was at the 95th percentile (if my memory is correct).

I slipped back a bit for a while and just clawed my back to 1493 last week. This time I was only 90th percentile. Now I am 1509 and 92nd percentile.

It seems that 1493 changed from 95th percentile to 90th percentile quite quickly. This is a big shift considering there are more than seven hundred thousand of players in Daily.

Is there any data where we can see the trend over time of the various percentiles?

A time lapse bar chart or distribution curve would be even better.

Is my memory correct or have I got that wrong about 1493 having changed by 5% in 2 months?

I am reluctant to rely on your memory. Why not take screenshots every now and then and make a graph?

My recollection (whether it is correct or not) just got me wondering on whether there are any stats on percentiles and how the may have varied over time.

e.g. Has the rating value associated with the 50th percentile for Daily or Rapid drifted up or down or stayed the same over the last 5 years? 10 years?

Just curious.

I notice that my daily rating is now 1545 and 90.4%

I definitely remember being well above the 90th percentile a while back at Christmas with a rating of 1493

This seems like a big change in a short time in the score required to get over the 90th percentile

I would be very curious to see how the median score  (and maybe upper and lower quartiles) for Daily (and the other time controls) has changed historically.

Is this info available anywhere?

Maybe we are witnessing an acceleration of the Flynn Effect on Chess.com!

mine, too, dropped about five. Sorry, I no screenshots.

Mine, too, dropped about five. Sorry, no screenshot.

While it's relatively straightforward to graph your rating over time, graphing the percentile over time would require looking up historical ratings for every single daily chess player on the site at the time.  This is highly unlikely to be practical.  If you're interested in it, you could track it yourself in a spreadsheet.

Yes to retroactively calculate it from the raw data would be quite complex (although not a huge computational task).

I was thinking maybe the guardians of the secrets of Chess.com might have kept a running track of the main measurements such as median and upper and lower quartiles hidden away somewhere in a secret global population performance log.

Anyway looking at Global for Daily today the population median appears to be:- Avg Rating - 1203.39

(is that what average rating means on the leader board panel on the left?)

So lets see where it is in a week. and then in a month...

I'm pretty sure they don't log historical percentile stats, at least not on a member-level basis.

No I did not expect they would log history for each member. I just thought they might have a historical trend log of the population median ranking (and maybe upper and lower quartiles?) over time.

Today my ranking was:-

17/07/2018 1533 Global (699,021) #66,336 Percentile 89.9%

So I am 40 points higher than at Christmas but I am now outside the 90th percentile.

It seems like the median (or at least the 90th percentile) must be increasing at a steady rate.

Very curious.

timer problem with network problem

My Daily rating is slowly climbing but my percentile rank is falling.

1547 Global (308,488) #39,792 Percentile 87.1%

The population median is obviously climbing at a steady rate of knots.

Is there any way to know what the median is for the population in Daily?

Find someone rated below you and check their percentile.  If they're above 50%, find someone lower.  If they're below, find someone higher.  When you find someone at 50%, their rating is the median.

It's probably about 1200-1250.

^Yes, I could indeed work it out from first principles.

Thanks for that MG!

I'd imagine the distribution of rating with respect to percentile is correlated with interest in chess.  I'm guess around 2018 when you posted this that chess website popularity gained traction, and the proportion of higher-rated players grew.  I'd bet $3.63 of your money the Netflix Revolution precipitated sharp percentile decline relative to rating.  I'm interested in whether this shift is more profound at lower or higher ratings.  Fascinating topic.