Are Ratings Accurate? (Original Research!)

Among all online chess platforms, is probably one of the most popular ones (the other being lichess) which spawns a debate whether its rating system is accurate? Is it accurate to over-the-board ratings or even better than them?

This is an interesting question that I will provide an answer for. Here is what I know on this topic: ratings are higher by about 100 points when compared to Fide ratings, on the other hand, it is only about 50 points higher when compared to USCF ratings. This only applies to rapid/classical rating and bullet/blitz ratings can reach 100-200 points higher than Fide.

Many players who have participated in online chess will come to at some point, this is a good question to answer since people can determine their overall strength with a good rating system.

Without further ado, let’s get started. ratings are slightly higher than Fide and USCF ratings

It is no wonder that ratings are naturally much higher than Fide and USCF standards only because online games are much more accessible than over-the-board games.

There are actually those who claim that ratings are much closer to real life, which is missing the bigger picture.

It may be true that the system by which is implemented would bring the best similarity to Fide standards, but the fact that people can play wherever they want makes the rating inaccurate.

In over-the-board games there usually isn’t as much opportunity to gain a huge rating, the games are limited after all.

In Fide, players have to visit international chess events with travel and accommodation expenses that would make every game much more meaningful.

There isn’t much opportunity to recover lost ratings for example, which makes Fide and USCF ratings naturally lower than that of’s.

Accounting for the inflation of ratings (online chess makes people naturally higher rated) it is warranted to think that ratings are some 100 points higher than the international Fide standard, and 50 points higher than the USCF standard.

USCF ratings (United states chess federation) are naturally much higher than that of Fide which makes the difference to only about 50 points.

This number is what is most agreed on by experts, but this is likely to only be an estimate and the actual difference would not be a rounded number (still, it is a good estimate to go by).

Different time controls may have more variance than real-life ratings

With platforms like, it is important to understand that their rating system usually affects faster time controls more than lower time controls. players vs. Lichess players. See who is stronger in this other article I wrote.

The cause of rating inflation in is the easy accessibility in the first place, the effect is further magnified in blitz or bullet games.

The classical time control within is usually much more accurate to real-life ratings since players have to actually play for long periods of time.

In rapid/classical it is a good guide to follow the 100 point estimate since the inflation would be minimized at least.

In blitz or bullet however the accuracy to Fide ratings could even reach 200 to 300 points, this is more than the estimated 100 point difference that could be applied in a normal rapid/classical rating.

People naturally play more games in faster time controls after all which makes the inflation more pronounced.

Different time controls are more or less different in their accuracy to real-life ratings (although the standard above generally applies).

Rapid and classical ratings are usually closer to over-the-board ratings while blitz and bullet one can be inflated by some +200 in some cases.

Many claims that acquisition of ratings are more difficult above 2000 Elo

This is another factor that needs to be taken into consideration, in over-the-board tournaments players are matched with all kinds of opponents that have varied ratings (sometimes even way lower than their own).

This makes over-the-board games quite easy in some instances since an opponent is a much lower-rated player.

Such a phenomenon can of course still happen on, however, their matchmaking system is better in finding opponents that have more or less equal strength.

This makes the climb beyond 2000+ Elo much harder in than with over-the-board ratings (as claimed by some at least).

There are many titled players (international masters and grandmasters) that express the difficulty of gaining a rating above 2000 Elo on

Some even say that after the 2000 mark, rating becomes more difficult to acquire than over the board ratings.

Whether the matchmaking system really makes it harder to acquire ratings beyond 2000 Elo is still in question, the accessibility might still even out the ratings.

The validity of this is still obscure and there are many titled players that claim the opposite, a further study into this needs to be implemented in order to get a full answer.

Fide and USCF ratings are still more accurate than ratings

There is an argument whether ratings are more accurate in determining a player’s competency when compared to Fide and USCF.

There is a substance to this and I personally believe that over the board (Fide and USCF) are still better at the moment.

Some people believe that the accessibility factor should be a non-factor in the first place, the limited accessibility restricts a player’s ability to show their full potential in Elo numbers.

This makes ratings more accurate than over-the-board ratings since players are able to recover from one bad game. 

This would be true if it weren’t for one variable that is present online, cheaters. There are many creative ways cheaters have found the game around the system, these unwarranted players throw the Elo rating pool in a complete mess.

Over the board games might pose limited accessibility however, the games are much more regulated, cheaters are almost always exposed. ratings in comparison seem superficial due to the existence of cheaters which makes the overall rating numbers inaccurate.

Due to the existence of cheaters, it seems that Fide and USCF are still the more accurate option.


The rating system is pretty good in determining a player’s overall strength (much better than lichess anyway) but the fact that it is played online poses a lot of issues.

The games are accessible which caused inflation on the numbers, plus cheaters have thrown the system into turmoil.

It is not accurate but it is the best that we have online, not a lot of people are willing to participate in over-the-board games and would be fine even with slightly inaccurate ratings.

I can say that a good rating will also have a good over-the-board equivalent, thank you for reading.