Upgrade to Chess.com Premium!

chess.com ratings are deflated against USCF


  • 22 months ago · Quote · #1

    AdamRinkleff

    So, I've read all the posts saying that you can't compare chess.com and uscf ratings, but that's obviously nonsense. If a person has two ratings, you most certainly can compare them. If a group of people have ratings, you can discuss them as a group.

    I've also read all the posts that say that chess.com ratings are inflated... uh, no.

    I've checked about about twenty people I know, who actively play USCF tournaments and chess.com blitz, with ratings ranging from 1200 (uscf) to 2300 (uscf). In -every- case, their blitz rating is about 2-300 points lower on chess.com when compared with their uscf standard rating. I think this is a pretty consistent pattern.

    Does this mean that someone with a chess.com blitz rating of 1500 is as good as a uscf 17-1800? no. long multi-hour games require a degree of discipline and concentration which most blitz players won't have. however, i think it is fairly clear that an 1800 uscf will normally be about 15-1600 on chess.com blitz.

    Should we be surprised that chess.com ratings are deflated? No, we already know from FIDE that the international pool of players is stronger than the American average. [Mod Edit: Please refrain from making offensive comments about Nationality.] That probably accounts for about 100-150 points. The rest is probably caused by the turbulent nature of blitz and differences in how chess.com calculates ratings.

    -Adam Rinkleff

  • 22 months ago · Quote · #2

    SmyslovFan

    Adam,

    • Chess.com ratings are calculated slightly differently than USCF ratings are.
    • Chess.com has a different population than USCF has.
    • Chess.com's "online rating" is based on correspondence chess, which is a different time control than USCF's. There is no time control that matches USCF's time controls.

    Any comparison between the two ratings can only be done on a very general basis, knowing that some players will have ratings vastly different from others.

    Ratings here are not deflated. They measure the population that plays.

  • 22 months ago · Quote · #3

    AdamRinkleff

    Smyslov, work on the reading comprehension skills:

    1) My post made it clear that I know chess.com ratings are calculated "slightly" differently, with slightly being a key word

    2) chess.com has a lot of the same people as USCF

    3) I'm talking about blitz, not online chess -- get a clue, this is the live chess forum

    4) any comparison expects some players to be exceptions, but... in general, there is a consistent 2-300 point difference between blitz here and uscf standard (your rating included sweetheart)

    5) ratings here are deflated, and if you read my original post, i acknowledge that the international population is better at chess than the american population, and this obviously serves to deflate ratings

  • 22 months ago · Quote · #4

    Sunshiny

    The problem is you're comparing blitz against standard ratings. I'm sure you must've noticed any one person on Chess.com can have multiple ratings that differ by hundreds of points. It shows that the person might be stronger in standard, but weaker in blitz.

  • 22 months ago · Quote · #5

    AdamRinkleff

    Sunshiny wrote:

    The problem is you're comparing blitz against standard ratings. I'm sure you must've noticed any one person on Chess.com can have multiple ratings that differ by hundreds of points. It shows that the person might be stronger in standard, but weaker in blitz.

    No, I think you are just making assumptions. If the two dozen USCF members I've examined have chess.com blitz ratings about 2-300 points lower than their standard uscf rating, then its a pretty decent hypothesis to conclude that experienced standard chess players have consistent performance at blitz.

    This makes sense to me. Just because you are running out of time, doesn't mean you forget how to do tactics. Most experienced standard players have already played thousands of games of blitz, so they aren't going to just suddenly forget how to play.

    It doesn't work the other way, however. Someone who is 1200 blitz is probably not going to walk into a USCF tournament and suddenly score at 1500. However, someone who is 1500 uscf, will probably be about 12-1300 blitz at chess.com. I'd bet money on it. There will always be exceptions, but this pattern is pretty obvious to anyone who compares USCF and chess.com ratings.

  • 22 months ago · Quote · #6

    ponz111

    Yes, Sunshiny, you got it right. You cannot compare USCF ratings which is tournament chess at a farly slow pace with a chess.com blitz rating. Apples and oranges as they say...

  • 22 months ago · Quote · #7

    AdamRinkleff

    ponz111 wrote:

    Yes, Sunshiny, you got it right. You cannot compare USCF ratings which is tournament chess at a farly slow pace with a chess.com blitz rating. Apples and oranges as they say...

    Ponz -- I have a USCF standard rating, and a chess.com blitz rating. Its not that difficult to compare them. ;o It's mostly a matter of subtracting one rating from the other.

  • 22 months ago · Quote · #8

    furtiveking

    If you honestly believe that you can draw accurate comparisons between ratings from two different chess organizations, you probably need to take a statistics class.

    In both cases, the ratings only measure one thing, the relative strength of a player against the other players in that organization. It's not a "true" chess rating, there is no such thing. So, if both ratings are a relative measure, there's no way to mathematically state how one relates to the other, since the populations of both organizations are different.

  • 22 months ago · Quote · #9

    AdamRinkleff

    Furtiveking, I've taken statistics, thanks though. I've taken lots of maths actually. You guys really don't know what you are talking about, which is why I thought I'd try to educate you.

    Yes, ratings measure the relative strength of a player against the player pool. No kidding? And when you have the same players, in two different pools, you CAN examine what the average difference is. This is not rocket science guys.

    Here is a chart to help you understand:

    USCF   Chess.com    Difference

    1800    1500             300
    2100    1850             250
    1200    1000             200
    1500    1275             225
    1750    1475             275
    1850    1575             275
    1400    1175             225
    1300    1100             200
    2000    1750             250
    1900    1600             300

    Average Difference: 250

    You guys can do this for yourself, if you are smart enough to handle the equations. I think you guys have been so busy believing there isn't any way to compare the rating pools, that you haven't quite considered that this -is- statistics, and there is always a mathematical relationship between any two sets of numbers.

    We are talking about comparing chess ratings with chess ratings... lol, that's not apples and oranges, that's chess and chess. Its the same thing. When you have the same people, playing the same game, trust me... there is going to be a mathematical relationship which describes their average ability.

  • 22 months ago · Quote · #10

    furtiveking

    First off, the pools don't have all of the same players. And 10 examples from a sample of 1000s is meaningless. If you truly understood statistics, you would know this.

    Just because there are some examples that support your theory doesn't make you right. Show me the differences between EVERY player in both organizations and we can talk, but until you can do that. Stick to playing chess and not trolling the forums.

  • 22 months ago · Quote · #11

    AdamRinkleff

    furtiveking wrote:

    First off, the pools don't have all of the same players. And 10 examples from a sample of 1000s is meaningless. If you truly understood statistics, you would know this.

    Actually, if -you- took statistics, maybe you'd know that you can extrapolate from a small subset. Isn't that the primary point of statistics, to deduce overall patterns without examining every single instance of a larger field? Can you even understand the words I'm writing? Hmm?

    Lets say you dropped a ball from a building, and you dropped it twenty-five times, and everytime you measured approximately the same rate of velocity. So, pretend you told someone what you measured, and they said, "Just because there are some examples that support your theory doesn't make you right. Show me the velocity of every ball that ever dropped and we can talk." Uh, yah, I'll get right on that... lol.

    Look, the thing about science, is you can duplicate experiments and verify the results for yourself. If you know people who actively play chess.com and uscf, it shouldn't be hard for you to compare the ratings. I know plenty of people who play weekly USCF tournaments and daily chess.com, and they all follow this basic pattern.

    USCF standard ratings tend to be about 250 points higher than chess.com blitz ratings. That's just a fact. Check it out for yourself. Its going to require that you learn how to add, subtract, and divide. Can you find an average? Check out wikipedia if you are having trouble.

  • 22 months ago · Quote · #12

    ElKitch

    AdamRinkleff wrote:

     

    -Adam Rinkleff

    I read your post and thought: "then where is he from?". I chuckled when hovered my mouse over it. 

  • 22 months ago · Quote · #13

    AdamRinkleff

    ElKitch wrote:

    I read your post and thought: "then where is he from?". I chuckled when hovered my mouse over it. 

    The great leader has a rating of 30000.

  • 22 months ago · Quote · #14

    furtiveking

    Sure you can extrapolate, but 10 examples isn't NEARLY enough in this case.

    That said, thanks for the laugh, I'm done feeding the troll.

  • 22 months ago · Quote · #15

    ponz111

    How can you line up two sets of ratings and compare when you are comparing two different games?  Speed chess is not the same as slow chess at all. Some people are great at speed chess and bad at slow chess.

    My speed chess rating is very poor as I have  disabilities which greatly hinder my performancee but my "slow" rating if I played would be more than 500 points higher.

    You cannot say chess = chess when the two chess variants are so different. So, again, I say apples and blues berries.... 

  • 22 months ago · Quote · #16

    AdamRinkleff

    furtiveking wrote:

    Sure you can extrapolate, but 10 examples isn't NEARLY enough in this case.

    Oh, honey, I know I said that I have more than 10. I think I have plenty. If you want to calculate the average ratings of a 1000 people, I don't think your ultimate average will vary considerably. I think its pretty clear that chess.com blitz ratings are about 2-300 below uscf standard ratings. Every active uscf player I know who uses chess.com has agreed with me that this is the case.

  • 22 months ago · Quote · #17

    ElKitch

    furtiveking wrote:

    If you honestly believe that you can draw accurate comparisons between ratings from two different chess organizations, you probably need to take a statistics class.

    I too think your rating mostly shows in what percentile of the group you are at. If one group is extremely strong (skewed to the right) then your a 1700 rating can be one of the lowest. And here 1700 is a solid and by most a feared player.

    But could it be that if two groups roughly have the same distribution, that the ratings are comparable for a large chunk of the group?

  • 22 months ago · Quote · #18

    AdamRinkleff

    My speed chess rating is very poor as I have  disabilities which greatly hinder my performancee but my "slow" rating if I played would be more than 500 points higher.

    Um, ok, so you have a disability. That is called a statistical aberration. You are an exception. How many people have your disability? When I'm talking about averages, your particular situation becomes irrelevant. You are an anomaly.

  • 22 months ago · Quote · #19

    DavidMertz1

    AdamRinkleff wrote:


    USCF standard ratings tend to be about 250 points higher than chess.com blitz ratings. That's just a fact. Check it out for yourself. 

    My blitz rating is around 1575 and my USCF rating is around 1400.  So, should I expect my Blitz raing to go down to 1150?  Or should I expect my USCF rating to go up to 1825?

    ... Somehow, I doubt it.

  • 22 months ago · Quote · #20

    AdamRinkleff

    ElKitch wrote:I too think your rating mostly shows in what percentile of the group you are at.

    That's exactly right. The rating shows your relative strength against the average strength of the rating pool. The chess.com blitz pool is about 250 points stronger than the uscf standard pool. This shouldn't surprise people, since chess.com is international, and incorporates a large number of people who play blitz on a daily basis. We should expect USCF standard players to experience a rating drop here. I just find it strange people talk about the ratings being inflated -- they aren't.


Back to Top

Post your reply: