Found a bug
"On average Chess.com: Blitz is 350256 points lower than Chess.com: Rapid"
The graph looks fine though
I built a website to compare chess ratings between Chess.com, FIDE, and USCF
Sort:
streetflame wrote:
Found a bug
"On average Chess.com: Blitz is 350256 points lower than Chess.com: Rapid"
The graph looks fine though
Yeah, I've been working on tyring to fix that bug. The actualy difference is 350 points, but there is a decimal missing. It happens only 1/20 times so its hard to re-create that bug when I want it to happen.
omanmenar wrote:
add the lichess and chess24 too
I didn't even realize chess24 was a place people played. Are there many people who play there?
Aug 5, 2021
0
#5
DavidDoesChess wrote:
omanmenar wrote:
add the lichess and chess24 too
I didn't even realize chess24 was a place people played. Are there many people who play there?
Yeah, the world champ plays there
Aug 5, 2021
0
#6
I think that this is a great idea 
My USCF is provisional though so I'm not gonna put it in as it might not be accurate
aMazeMove 写道:
DavidDoesChess wrote:
omanmenar wrote:
add the lichess and chess24 too
I didn't even realize chess24 was a place people played. Are there many people who play there?
Yeah, the world champ plays there
Ah so there are many world champions
Wait...
Aug 5, 2021
0
#8
Nice website but i think it would be a lot better if you get a legit data set. People can enter whatever they want in the form and it would skew the results.
Aug 6, 2021
0
#9
aMazeMove wrote:
DavidDoesChess wrote:
omanmenar wrote:
add the lichess and chess24 too
I didn't even realize chess24 was a place people played. Are there many people who play there?
Yeah, the world champ plays there
That’s pretty much it. I tried to play there once. I never got a game. People only play bullet there.
kingattacker3 wrote:
I put in my data. Happy to contribute, as I am curious as we get more sample points what the results would show. Good job! Thanks.
Thanks for the kind words and for contributing!
What an interesting and cool website, very nice and well done. For example interesting how at the lower ratings lichess is inflated compared to chess.com blitz while at the very high ratings they seem to become almost equal.
One thing you could do is maybe you could ask if they have played within something like the past year or if they think it's a good representation of their strength as for example you might have an account but not played there since you were a beginner or you might have a FIDE account but not been playing rated matches for a long time despite having changed strength.
You could also ask people not to enter their ratings again - there are a lot of very stupid people out there, remember there are some very young children as well. You should also probably track their IPs and cookies for jokers or trolls also - again, obviously why would someone be so ridiculous as to lie on a site like that and what purpose would it serve but it's amazing how mentally ill some people are out there and many of them seem to gravitate to chess, if you witnessed some of the trolling on this site it's amazing what some people will pull. Maybe you could even ask them to put their handles on the different sites and maybe even real name if they were willing (not public) so you could check out the data better. Keep in mind also you will probably have to comply with GDPR for dealing with data from the EU. It's probably fine as it is, but just to try to validate the results a bit more, especially regarding inputting inactive or very old ratings.
horselover123 wrote:
Hm, should I exclude any incredibly skewed ratings, or should I just leave it
(my underrated uscf, my puzzles which have been resetted a million times and are far from their peak, my daily that i rage-quitted 15 games in, etc.)
Btw, I encountered that bug again, on average USCF 1052197 points lower than chesscom rapid
If you don't think its a good representative of your skill level it would make sense to leave it out, but if people do leave in numbers they don't think are a good representation there will hopefully be enough data for each pair that it won't throw off the numbers by that much
Uhohspaghettio1 wrote:
What an interesting and cool website, very nice and well done. For example interesting how at the lower ratings lichess is inflated compared to chess.com blitz while at the very high ratings they seem to become almost equal.
One thing you could do is maybe you could ask if they have played within something like the past year or if they think it's a good representation of their strength as for example you might have an account but not played there since you were a beginner or you might have a FIDE account but not been playing rated matches for a long time despite having changed strength.
You could also ask people not to enter their ratings again - there are a lot of very stupid people out there, remember there are some very young children as well. You should also probably track their IPs and cookies for jokers or trolls also - again, obviously why would someone be so ridiculous as to lie on a site like that and what purpose would it serve but it's amazing how mentally ill some people are out there and many of them seem to gravitate to chess, if you witnessed some of the trolling on this site it's amazing what some people will pull. Maybe you could even ask them to put their handles on the different sites and maybe even real name if they were willing (not public) so you could check out the data better. Keep in mind also you will probably have to comply with GDPR for dealing with data from the EU. It's probably fine as it is, but just to try to validate the results a bit more, especially regarding inputting inactive or very old ratings.
I am just now reading this, but you make a really good point. I've gone both ways about adding a cookie to keep people from entering in data multiple times, so far, I've decided to allow people to add contribute multiple times because someone could come back in six months with different ratings, and that data would be beneficial. I also still have to learn about how to use cookies before I could implement that too XD
I feel like you should use median, because someone with like 30000 rating on chess.com puzzles and about 1900 on everything else could totally wreck the graph
ChesswithNickolay wrote:
It just says "thank you for your submission"
If you click the graphs link in the banner it will take you to a section where you can select what ratings you want to compare, but I love your feedback. You're the second person to have that thought and so I think I should add that to the feedback once you submit your ratings.
Is it working? I put in my data and checked the graph, but had to enter my data manually (EDIT: on the graph page) to get a comparison, and that was .. odd ..
"2330 chess.com puzzles would equal 154 chess.com rapid with a variation of 1258"
If I just go in again without entering my data first it says 2330 puzzles would equal 1523 rapid with a 201 variation
During the covid shutdowns I spent a lot of time playing chess and increased my online rating a good bit. Shortly after I was vaccinated I visited my local chess club. I met some new people and they soon asked me my rating. I didn't think my USCF rating was an accurate representation of my skill level, so I told them my chess.com rating instead. Which begs the question: "how do ratings from chess.com, USCF, and FIDE compare?"
So, I built a website to help answer that question:
www.chessratingcomparison.com
If you could give it a visit and fill out the form, I would really appreciate it. The website uses data from visitors to create its data-set. You can also take a look at the graphs section to see how different ratings compare. Also, this site is still a work in progress so if you encounter any bugs or have any feedback, please let me know!