FIDE rating distribution

Hi Keyser.  Have you succeded with your chart?  I was thinking you are better off not using Excel but just do a short code in any language like Fortran or C++.  The code gives you the advantage of unlimitted players and basically all what the code would do is count players with such and such rating and then give you a file that you can plot in Excel if you like.  Probably if any difficulty at all would be in actually reading your long FIDE file unless you know well about formats and stuff.  If you get on with the project in Fortran I can give a hand with the coding.  Cheers.

That's the thing I did :) To say the truth, jnieuwen's Unix way is much better. He did not invent bicycles.

Hey KahZeeMin.  Yes I can see that Unix post now.  I am actually a PC user, no server, no UNIX, no Linux either.  I was just offering another solution to Keyser.  Nothing wrong with building my own bicycle.

So the average player is above 2k?

Goodnight sweet selfconfidence.

@LikeTheLake

There's nothing wrong. But I do have Linux on my desktop and I did write an app to get this graph. I was to figure out an easier way. My bad, I have to learn more about the OS I'm using.

@Naraga

Acording to this graph 50 percentile is between 1900 and 1950, so average FIDE-rated player is a bit below 2k. A FIDE-rated player, you know. Professional of a kind. AFAIK, average club player is rated about 1500, so both of us are OK :)

@KahZeeMin Well, I also cheated in a certain way, the script to convert the csv to a graph I had already laying around :)

The FIDE rating is atypical, as for a long time, you needed a minimum of 2200 rating to be listed as FIDE. Then this was progressively lowered to 2000, 1800, etc. and now as low as 1000.

This explains why the average FIDE rating is still high, but it's bound to slide down in the coming years, as most new players now enter the charts U2000.

Thanks for your help guys, you did a great job. Sorry for delay, some computer problems

I downloaded the recent FIDE list and did some quick statistics.

The # of active rated players is rising (99 611 vs. 90 271 last year, roughly +10%), but the average rating of rated players is down 30 pts. from 1895,1 to 1862,7 so there's significant rating deflation in the pool, as expected.

Some additional stats :

Top 1% elo cut-off : 2477 (last year : 2488)

Top 10% elo : 2206 (last year 2223)

Top 50% elo : 1877 (last year 1907)

Here is the updated rating distribution chart : https://docs.google.com/drawings/d/1koiKWVmyl4ULBypAyEcvLWrZiVc79swcwbNKpvEaTbs/edit?usp=sharing

If you want to do your own processing of the data presented in the OP you can use the following perl script ( note that perl is the best language for such data processing jobs and it is free ):

use POSIX;

open(DIST,"players.xml");

while(($line=<DIST>)&&($no_players<1000000))

{

#print $line;

if($line=~/<player>/)

{

$no_players++;

print "--- processing player $no_players
";

while(($line=<DIST>)&&(!($line=~/<\/player>/)))

{

#print $line;

if($line=~/<rating>([^<]+)/)

{

$no_ratings++;

$rating=$1;

print "--- rating $no_ratings = $rating
";

$rating_100=floor($rating/100);

$dist{$rating_100}++;

}

}

}

}

close(DIST);

open(DIST,">dist.txt");

foreach(sort(keys(%dist)))

{

$key=$_;

print DIST $key,"00\t$dist{$key}
";

}

close(DIST);

I have actually run the script and got the following result:

The distribution is a tab delimited text file ( dist.txt ) which can be copy pasted into an Open office ( also free program ) sheet. You can easily create a chart from it:

Distribution of rapid ratings:

Distribution of blitz ratings:

Here is a quick comparison with last year rating distribution :

https://docs.google.com/drawings/d/1Eeu-xMNWtx6-_A08tUn5sGmHwBPZwpIKdzEB4NCJnYY/edit?usp=sharing

This is a very good point.  The GM training me lives in Serbia, and he's mentioned that there are tons of isolated local players who don't know much chess but who are 2000-2200 because FIDE only used to award players with ratings above 2000 (and 2200 before that).  For that reason, FIDE ratings can be goofy, when viewed collectively and globally.  I am sure the ratings are a little more sane and suited to the player's strength at events that aren't simply a bunch casual local players playing.

As of 14MAR2016, median rating (world, standard) is now 1822, average rating 1794.
Data from http://ratings.fide.com/download.phtml

@hicetnunc the fall continues... 

Thank you for this update