Major Proposed Change: Ratings Cap Above 2400 for Unverified Players

no, we haven't announced it. also it's not true. here, let me officially announce for you: "NOT all non-premium members over 2400 are cheating."

Because I am silly and derive a sense of accomplishment from seeing my rating go up (on the rare occasions that happens :-P ) - it does, after all, reflect my strength to some degree.  Because this, in my view, crosses the line on what is acceptable to withhold from nonpaying members, and I've grown attached to chess.com and would resent that kind of treatment from a site I really like.

I realize that something needs to be done about cheaters, but these small things really matter to me, and they're enough to make me resist this change until I'm convinced there's no other way to stop repeat offenders.  At least give the community some time to discuss this.

(Also, do you think members under 2100 or so will be affected by this either way?  I only ask because there's been "ban the cheaters at all costs" talk from people who might not be impacted at all.)

Maybe you could employ someone with the strategy of only them doing a job that involves looking for cheaters. This means they would be able to dedicate all their time to this. I don't know if this would work but sure you could look into this. I would be happy to do the job. I am sure everyone at chess.com has ideas on how to try and stop this too who could pitch in. As for your main idea its fair on some but not others in saying that its not much a year to pay and it comes with perks which i think are more than worth it.

Why not just a cap of 2400 for the first 6 months? After that, a player obviously has some form of loyalty to the site.

(EDIT: I guess shoop2 already said this.)

it's an interesting idea dunce, but i see two problems:

1) we will still have to investigate them for cheating when they cross 2400, which will still cost us money.

2) i have thought of a way for a cheater to circumvent that idea. i won't say it here because if we ever did something like this i don't want to help come to the idea.

The 2400 cap for non-paying members is a VERY good idea. In fact it acts as a penalty: if you cheat to get high ratings, you are in danger of losing your money!

I wonder what would happen to paying members that get over that point and then stop paying. Their rating would just freeze or it would change to 2400?

nxavar, it would revert to 2400.

 In that case I think there should be some kind of "memory". That is, when he starts paying again the old rating would be restored. There's also the case of someone missing to "add credits to his account" while having auto-renew off and various other cases of missing the payment (such as a maxed-out credit card at the time of renewal).

This may not be workable, but how about having an automated program that takes only 1 or 2 PGN files (not randomly, only credible games, a detail to be worked out if this is even a good idea) and checks for how many moves match with a computer.  If the match % is high enough then that user gets flaged and gets a rating cap (say 2400 or whatever).

Later if they're flagged again, maybe it takes 5 games for the 2nd offense, then they recieve a lower cap or are banned etc.

I suppose if automated checking were feasible it would have already been done...  I like the loyalty idea though that dunce and shoop proposed.

Paying members can exceed 2400, but non-paying members can't?

It's too bad gbidari got his CM title.  It would have been fun to read the angry posts from him.

agree nxavar, we would want to save the rating.

DeepGreene wrote that a player would be given the option to either upgrade their membership or confirm their chess title. What method did you have in mind for players' "confirm[ing] their chess title"?

If every player who crossed the 2400 mark had to play a mandatory 10-game set against the Chess.com computer, that should provide enough data, no?

that is already in place dunce, we've been confirming titled players accounts for years now. they just have to contact us.

Does that hold true for people who don't have official titles, but have high ratings?

if they don't have an official title, then they don't have any title that we could confirm. am i misunderstanding your question?

Probably my misunderstanding . . . You confirm that they actually hold a title, or you confirm their rating by playing against them with a known-strength player/engine?

PS, I think I edited my post #38 since you responded.

Hmm, I think the problem is you can't completely automate it... it's hard for a computer to estimate a player's strength.  But if it takes staff time to review they'd need to hire a bunch more people so I don't think this would work.

(Plus they could just cheat vs the engine by using an engine themselves :)

for titled players we confirm that they actually hold a title.

we *also* will still investigate them for cheating for various reasons, but that's a separate matter.

you see the "im" next to my user name? anyone with one of those titles next to their name, like "nm" "fm" "im" "gm" etc. that means that we have confirmed their identity and that they have a title. (however if they wish their real name remains known only to chess.com staff, as opposed to public knowledge).

You may be right.

However, as to the cheating using an engine, I believe that is fairly easy to sniff out.

And as far as an engine's accurate assessment of a player's ability, Fritz and Rybka (and probably others) can create a handicap mode where they can play at your level. So I assume they can generate an approximate estimate of your skill level.

it won't work, dunce :( but thanks for the suggestion.