Online ratings

[John Coleman]

How do online ratings, for example chess.com, compare with over-the-board ratings?

[Jack Maguire]

Re: Online ratings

I'm surprised that no one has responded to your query as yet, John. My own standard 'live' rating at chess.com is generally plus or minus 100 points of my CFC rating so I would most certainly say that the two are relatively tantamount.

[Felix Dumont]

Re: Online ratings

There's pretty much no correlation. I know some people who have an on the board rating of 1200 and an online rating (say on Chess.com) of 2100. Some national masters have ratings as low as 1800 or 1900. We often meet new players at chess clubs who have a rating on Chess.com of 1500 or more, and they lose very quickly to players who have an on the board rating of 1100 (but an online rating much lower than 1500).

We can't really compare ratings for two completely different time controls. Also, many chess websites have radically different rating systems.

I rarely play online, so I can't really tell you the differences between all the websites. I think Chess.com ratings might be slightly lower than CFC ratings for many players (except maybe for kids).

[Mathieu Cloutier]

Re: Online ratings

Well, the major factor here will be that most of the online activity is blitz/rapid games or even faster. Some players develop only their blitz/rapid skills that way. Agressive openings, a flair for quick attacks and 2 moves combinations. But to be succesful at standard time controls, you need a little more than that.

Inversely, some players are much better with slower time controls and can't manage playing a game in 5 minutes.

So in the end, you may have large difference between online and standard, OTB ratings. I've said it before, but for me it's just not the same game.

[Jack Maguire]

Re: Online ratings

FIDE now provides ratings at 3 different time controls and there's clearly a strong correlation between all 3. The same names tend to dominate all 3 categories. The CFC currently provides 2 different time ratings, 'Rating' and 'Active Rating'. Again, there's a strong correlation. My own differential between the two is a mere 8 points.

The most common time control in a 'standard' rated game at chess.com is 15/10 which means that a 60-move game gives each side 25 minutes, which is an excellent match to our own CFC 'Active Rating'. And from my own experience, the facts indicate that the two should be within a 100 or so points of one another.

I hope Bill Evans doesn't mind me bringing up his name, but he's the one fellow ACC player whom I know plays with at least some regularity at chess.com. And the empirical evidence therein duplicates my own musings. Bill currently has a 1977 'Standard' rating at chess.com which is just 21 points below his CFC 'Active Rating' of 1998 and 62 points below Bill's CFC 'Rating' of 2039.

I can't speak for other online rating systems, having only played at chess.com, but there would certainly seem to be a strong correlation between chess.com's 'Standard' rating and CFC rating, 'Active' or otherwise.

[Ed Seedhouse]

Re: Online ratings

Theoretically there is no reason why an Elo rating in one independent rating pool should have any relationship to an Elo rating in another pool. The ONLY thing that can be used in an Elo rating to predict results is the difference between ratings. The actual numerical rating is irrelevant for this, only the difference between two ratings in the same rating pool. (The same thing is true of Gliko ratings which many sites use instead of Elo.)

My online ratings for slow games varies between 1600 and 2060 (the latter being close to my five year old CFC rating) depending on the server I am playing on. For quick games it varies even more widely. This is complicated by the fact that many sites don't do even Elo ratings properly.

In short you really shouldn't compare your rating on any particular online site to any other, or to your CFC or USCF rating. Easier said than done, of course, and I tend do to avoid sites that give me ratings in the 1500's. But this is of course irrational and based on emotion, not logic.

[Bindi Cheng]

Re: Online ratings

I get all of my students to do 5 0 or 15/0 on ICC or 15/10 on chess.com. I think it has some correlation, obviously not +1 territory but I would say 0.5 area. I find that people who have higher ratings in blitz/standard time controls perform better in tournament play and are more consistent. I've had some students in the past who have had awful blitz ratings and it shows in their actual games - they might have a much higher over the board rating but they have some terrible games and it takes them a while to get used to it again. Overall I'm a big fan of using blitz ratings to figure out what a person's play is like and how comfortable they may be in real tournament play.

As a cautionary measure, I would like to say that those who play blitz all day long and not much else will not see much improvement in their game, just like those who study all day long and never get any blitz games in won't see much improvement either. It's vital to find the right balance.

[Roger Patterson]

Re: Online ratings

FIDE now provides ratings at 3 different time controls and there's clearly a strong correlation between all 3. The same names tend to dominate all 3 categories. The CFC currently provides 2 different time ratings, 'Rating' and 'Active Rating'. Again, there's a strong correlation. My own differential between the two is a mere 8 points.

The most common time control in a 'standard' rated game at chess.com is 15/10 which means that a 60-move game gives each side 25 minutes, which is an excellent match to our own CFC 'Active Rating'. And from my own experience, the facts indicate that the two should be within a 100 or so points of one another.

I hope Bill Evans doesn't mind me bringing up his name, but he's the one fellow ACC player whom I know plays with at least some regularity at chess.com. And the empirical evidence therein duplicates my own musings. Bill currently has a 1977 'Standard' rating at chess.com which is just 21 points below his CFC 'Active Rating' of 1998 and 62 points below Bill's CFC 'Rating' of 2039.

I can't speak for other online rating systems, having only played at chess.com, but there would certainly seem to be a strong correlation between chess.com's 'Standard' rating and CFC rating, 'Active' or otherwise.

For most people, their CFC active rating was set equal to their CFC rating a few years ago. Most people don't play many (or any) CFC active rated games so any correlation is between their current CFC rating and what it was a few years ago. (and not surprisingly, they are correlated)

[Dan Scoones]

Re: Online ratings

The main value of an online rating is that it lets you get interesting games against players who are at your own level. The number itself isn't that important. Of course, your rating and $2.00 is still good for a tall Pike at Starbucks.

[Digeng Du]

Re: Online ratings

had a higher ICC blitz rating many years ago when all I played was the KIA and KID. My moves were all bad but I sure played them quickly!

Today, after many hours watching the youtube games of chessexplained and gymbuddy, I can now proudly say not all my moves can be considered bad. But alas, although i have gained in knowlefge, old age has robbed my hands of their former speed and nimbleness...

[Tony Li]

Re: Online ratings

My Blitz rating is 300 points lower than CFC...

• I am not slow.
• But blitz players are extremely good at openings.
• Plus they don't respond to ghost threats since there is much less at stake.
• And they don't get lazy/complacent.

I find this is very instructive, I get free CFC rating points from younger players being lazy/complacent at the board and older players responding to ghost threats. Since the former group will grow with age, I would strongly recommend to older players to adapt to ghost threats.