Adding a maximum rating

Post by **gmiller** » Wed Oct 22, 2003 1:07 pm

Due to the number of obscenely high ratings (even those obtained legitimately), I'm thinking of making an absolute maximum rating. Say 2700, 2500? So no one's rating will ever go above that.

hamot · Post by **hamot** » Wed Oct 22, 2003 1:10 pm

Sounds like a good idea, since even many professionals never even get there.

keithstuart · Post by **keithstuart** » Wed Oct 22, 2003 1:45 pm

Sounds like a very good idea

I wonder is it possible to do something similar but run of the average rating of all opponents played as you could play people well below your rating 99% of the time but every time you play that 1% of higher rated players a win psuhes your rating up

jollyjumper · Post by **jollyjumper** » Wed Oct 22, 2003 3:12 pm

Hi,

IMHO an absolute max. rating wouldnt help the problem. To players from another sites it would look to be more sensible (maybe), but we all would be still overrated. My rating here is 3067. On ICCF it is 2217.

IMHO my net-chess-rating has to be multiplied by 0,72. Maybe other players here come to another multiplier, but then we could take an average.

On the other hand I think, the first problem that has to be solved is, why so much players have so high ratings. IMHO the reason is that everybody can register lots of user-IDs rated 2000, then playing against oneself and pushing up ones starting-rating to 2400. If new players start with an rating = 800 this couldnt happen anymore. It would be possible to push up ones rating to lets say 1200, but nobody has to care about that, because most players play better than 1200.

so long
Joerg

keithstuart · Post by **keithstuart** » Wed Oct 22, 2003 5:00 pm

There is always the exceptionally unpopular and highly radical step of everybody has their rating reset to either the same level or one of several lower levels depending on where their curretn rating is

ie

a single rating of 1200

or ratings relating to levels with new starting ratings of

800, 1000, 1200, 1400, 1600, and 1800

keithstuart · Post by **keithstuart** » Wed Oct 22, 2003 5:01 pm

Should of added all new players start at either 1200 or 800 depending on which system was implimented unless they can show a rating from another reputable source ie either USCF, FIDE, ICCF, IECG or BCF

davidswhite · Post by **davidswhite** » Thu Oct 23, 2003 12:34 am

Greg, I'd like to make the following suggestions:
all established ratings of 3000 and higher would be given a rating of 2600;
all other established ratings would be pro-rated accordingly(i.e. a rating
of 2700 would convert to 90% of 2600, or 2340);
individual ratings would only increase by wins or draws against higher-rated players or wins against same-rated players.
I haven't addressed the question of how to deal with provisionally-rated
players but I do think this would be a superbly appropriate time to
discontinue the use of rating floors.

Whatever you decide to do about ratings will be fine with me although I
believe that implementing the above suggestions(or something like them)
would go a long way to bring the ratings back down to earth without
instituting an absolute rating ceiling.

Just my 2 cents,
Best regards,David

tellymetwise · Post by **tellymetwise** » Thu Oct 23, 2003 5:18 am

Have you made a list of all possible reasons why net-chess ratings are inflated?

I've seen:
- the floor (%?) ex. floor of 2000 while playing at 1500
- cheating (%?) ex. creating fake accounts
- range of opponents rating (%?) ex. playing weak or too strong players
- too high starting rating (%?) ex. 9999 instead of 1200
- untimely timeouts (%?)
- (maybe) selective opponents challenge
- (maybe) some other reasons .....

NB.
- installing a ceiling requires a removed floor anyway.
- any change (ceiling/floor) will take slow effect due to CC nature.
- at best, a difference of 400 should show 90% win chance
How many steps of 400 are there between lowest to highest?
800 - 2800 suggests 5 steps of 400 or ....
A wins 90% against B, B wins 90% against C etc, etc)

Maybe some other suggestions can give more insight in what you want to change for a long time durable effect?

katchum · Post by **katchum** » Thu Oct 23, 2003 6:21 am

If you play on an online chess applet, ratings don't have rating floors and you will not time out due to vacations and such, you cannot cheat, only if you use two computers against yourself.

There are no rating ceilings and floors.
I'm happy.

peterpawn · Post by **peterpawn** » Thu Oct 23, 2003 8:49 am

Another cure, would be, to just stop putting so much stock into one's rating. It's only a guide to one's strength as a chess player. It's not absolute truth! There will always be high rated players that are easy to beat, and low rated one's that will wipe you out, your family, and everyone you know.

katchum · Post by **katchum** » Thu Oct 23, 2003 12:35 pm

On yahoo games I don't even get to 1600 while I easily get to 2000 on this site...

keithstuart · Post by **keithstuart** » Thu Oct 23, 2003 1:54 pm

katchum wrote:On yahoo games I don't even get to 1600 while I easily get to 2000 on this site...

You should of seen my rating at Blitz Chess on ICC it was in 3 figures

keithstuart · Post by **keithstuart** » Wed Nov 05, 2003 8:01 am

I had a thought just know do we actually have a maximum figure for the rating floor??

As otherwise with careful game management you can slowly keep on rising forever

If we do have one what is it?

And also prhaps reducing it down to a lower level like 2000 might help

katchum · Post by **katchum** » Wed Nov 05, 2003 10:02 am

I guess it's 2300

knightmare · Post by **knightmare** » Wed Nov 05, 2003 6:19 pm

Instead of a Maximum rating, I think it would be much more useful to come up with a formula that would kept the AVERAGE rating at around 1500. This would make everyones rating more meaningful to the general chess community.

tellymetwise · Post by **tellymetwise** » Thu Nov 06, 2003 8:32 am

What I noticed with net-chess is that wins due to timeouts can increase ones rating, the decreasing of the loser rating is stopped by his/her floor rating.

If a weaker player timesout, this shouldn't provide any problems as the weaker player will probably loose more often then not.

But if a stronger player timeouts, as keith mentioned he did a couple of times, then inflation of ratings will occur. With a raised floor, keith (only as example) now can raise his own rating based on this floor rating as his relative strength will be recalculated on this "to high" floor rating.

The same applies for players of equal strength.

Maybe masss timeouts (or resignations as with hsingletary vs keith) can handled (or avoided) in a different way?

FYI,

According to USCF: (quick link to pdf)

RATING FLOORS
Rating floors exist at 100, 1400, 1500, 1600, ..., 2100.
No player's rating can drop below 100. A player's rating
floor is calculated by subtracting 200 points from
the highest attained established rating, and then
using the floor just below. For example, if a player's
highest rating was 1941, then subtracting 200 yields
1741, and the floor just below is 1700. Thus the player's
rating cannot go below 1700. If a player's highest
rating was 1588, then subtracting 200 yields 1388,
and the next lowest floor is 100, which is this player's
floor.

Post by **gmiller** » Thu Nov 06, 2003 11:51 am

Thanks for the link, I haven't read it yet, but will try to make the rating calculations here as much like it as possible.

Net-Chess.Comhttps://www.net-chess.com/phpBB2/

Adding a maximum rating

Adding a maximum rating

There is only this way

yes

Re: yes