I created a website to compare chess ratings

how will it deal with people putting fake/wrong data? that might be a problem

Yes but why would they?
~ Even if they do, standard deviation and variance might ease the estimation with the fact that most of them will put true data and such kind of people are less.

Thanks @hor@Akbar2thegreat said in #9:
> @DavidDoesChess
> For me, a statistics lover, this was interesting filling out the form.
> I have seen many asking their OTB rating based on their rating at online chess.
> Hope to see something like chessgoals website.

Thanks! I'm hoping to add a hover to the graph today where you can read the value for each point and the regression line (I actually had that working earlier and then it broke.) In the next few weeks I'm also hoping to add a page where you see the overall data as well.

@vellarco said in #11:
> how will it deal with people putting fake/wrong data? that might be a problem

Right now I can go into the postgres SQL database and delete individual records. It does take some time but I'm not sure of a good way to code something that would

@Finalstrenght said in #14:
> Why puzzle rating elo must be max 3200?

I put in hard limits so someone couldn't throw off the data spectacularly. Are there many people with a rating higher than 3200 puzzle here at Lichess? I'd imagine that would be the highest average ratings of all the ratings I ask for.

For random inputs by people in the form, you don't need to delete individual entries rather you can simply put conditional statement in SQL to kkep only logical values at least -
DELETE FROM <table_name> WHERE rating >3000 and rating< 600;
- With the fact that online rating is floored to a minimum of 600 and that someone like Zhigalko or Magnus won't fill but many others would. This would prevent unnecessary high and low rating that would deviate the values by a greater magnitude.
And a suggestion: Put the option of writing Lichess/ username to check them individually when the rating is near 600 or 3000.
I used this website to learn SQL while learning it in school 4 months ago.
And by the way, I recently learnt Interfacing of Python (which I learnt last year) with MySQL.

@FC-in-the-UK said in #19:
> I tried comparing blitz and bullet Lichess ratings.
> On average Rapid is 1437807 points higher than Bullet
> That doesn't sound right.

Yeah... I've gotten a similar numbers a few times, but each time I try to re-create the issue after I see it, it goes right away. I do hope to solve that soon. Right now I'm trying to fix a sneakier, but problematic issue. Right now when you look at the graph your browser is creating the graph over, and over, and over again as quickly as it can. Which is causing other issues.

This topic has been archived and can no longer be replied to.