I found a interesting pattern in the ratings history for an opponent I played last month, id# 2652981.
He has an pattern of losing lots of ratings points in G/30 and action tournaments until is rating is less than 1800. Then he plays well in the U1800 in the World Open and Atlantic Open. If his rating goes above 1800 he will then play in more short tournaments where he will, in general, do very poorly. He can play well in fast tournaments. In Jan 2004 he had a >2000 performance in one G/30 Quad. Otherwise, his record is 0-9-1 in these one day tournaments.
What do you think?
Mike Regan
2652981 is not a valid member number.
12652981 works however and I don’t think a 100 point deviation is that bad. I could go down 100 points in one bad tournament lol
Yep, I did the cut and paste wrong. The correct id is 12652981. Sure we can all lose 100 points. He just seems to have bad tournaments (as in winning no games) when his rating is over 1800 and he is playing in one day cheap tournaments. When he is playing in tournaments where there is a significant prize fund he does much better.
For example, in his two one day tournaments this spring he was 0-5-1 with a performance rating of 1220. At the World Open his performance rating was 1873. At the MD Open his performance rating was 1803.
So is it normal for a performance rating to vary so much?
Mike Regan
What do you mean by ‘normal’?
Are there numerous examples of players whose rating goes up and down? Yes. The most active player in the USCF, Jay Bonin, has a rating that fluctuates in a range from about 2350 to 2450.
Why is a much different question, and one hard to determine a precise answer to. It could be manipulation of the ratings system or it could be a player who is much better at slow chess than at faster time controls (or vice versa), or when there is serious money at stake, or because he plays better on weekends than on Tuesday nights, because the events draw from different pools of players, etc.
If his performance rating was 1873, then his sandbagging, if that’s what it was, was not successful. It would take a MUCH higher performance rating than that to win the under-2000 section of the World Open.
Bill Smythe
But he might be hard-of-hearing.
His 1873 performance was good enough for 6-3. He lost in the last round. If he had won, he would have tied for 5th place ($1230).
He appears to be a true high 1800 player (when he is playing for big money). So keeping his rating under 1800 does not ensure he will win u1800 sections. But ,it sure is easier than trying to win money in U2000 sections.
Mike - Yes, maybe he is bad a fast time controls. Why then does he keep playing fast time control tournaments? Personally, if I was 600 points worse at fast time controls, I wouldn’t play.
The prototype pattern for manipulating your rating is when it goes down in cheap short tournaments and goes up when there is money on the line. If when we see this pattern we explain it away, what stops someone from doing it?
Mike Regan
ps He didn’t need to be 1800 to beat me in our game. A 1500 player could have been across the board and won, the way I played.
I certainly cannot draw any conclusions about any particular individual. But in general I think Michael M. is making a fair point.
Michael M.'s point is not that the player’s rating fluctuates. Michael’s point is that those fluctuations mathematically correlate with prize money much more than can plausibly be accounted for by random chance.
It would be a violation of probability theory if such oddities never happened, so each individual can hide behind that for defense. A few can do so legitimately.
But if the pattern continues even after it has been noticed, well then at some point it becomes like the judge says: the standard is “beyond all reasonable doubt”, not “beyond all possible doubt”. Then a player could be required to play in the next higher brackett.
Maybe the easiest fix would be for T.O.'s to use each player’s “highest rating at any time during the past 12 months”, instead of their current rating.
Personally I suspect sandbagging is the biggest form of cheating; despite all the excited talk about tiny electronics. The possible technique that motivated the start of this thread is particularly effective. It would be more dull to purposely loose a 3 hour game than to throw a 40 minute game.
Money payouts are not the only this distorted by sandbaggers. The ratings of others are affected in a causal chain.
[size=134]**** TOURNAMENT ORGANIZERS ARE PARTLY TO BLAME ****[/size]
The USCF has labeled certain boundaries as classes A,B,C,D etc.
But no way does that mean T.O.'s should reuse those same boundaries over and over for pairings and prize funds!
Instead, T.O.'s should shuffle the boundaries each rerun of the tourney, for example:
2000 ,1800 ,1600 ,1400 ,1200 This year.
2100, 1900 ,1700 ,1500 ,1300 Next year.
I am sure many internet chess players cheat by getting help from a chess engine (as has been written of in Chess Life). But those games are not USCF rated.
Are the forums the appropriate place to make accusations such as this?
If you have evidence, file an ethics complaint. Otherwise, it’s not fair to drag someone through the mud when you only have conjecture and they aren’t around to defend themself.
What I meant to say was, it would take a much higher performance rating than 1873 to win the under-1800 section.
Bill Smythe
I like this idea. My first thought was why does it have to even be even hundreds? Why not groups of 2050, 1950 etc or 2025, 1925 etc
Then I thought why not define your prize groups by every 8 players irregardless of where the rating falls? The open sections would comprise the top 8 players and everyone would be eligable for this prize. But then your class or group prizes would be every X number of players with the bottom group being variable, with the prize amount adjusted to its’ size.
That or your group prizes would dependent on what number divided easily into the whole group.
The disadvantage here is there is no easy way to know ahead of time what group you will be playing in.
THe advantage is that there is no way to know ahead of time what group you might be in so it would be harder to sandbag for a specific tournament.
There is still nothing to keep someone from delibertlly getting his or her rating as low as possible, but the floors should help with that.
The player under discussion was assigned a CCA minimum rating of 1800 in mid-July.
Bill Goichberg
How about the Floor Class Championships?
U2001, U1901, U1801, U1701, etc. You’d only be eligible if you’ve been near or on your floor for at least 2 years. 
The Cat in the Hat’s Tournament Record and Prize Winnings (about $100,000).
1992
1992 MANHATTAN CC APRIL RATED BEG T 9th place/16
1992 World Open 215th U1600
1992 MARSHALL CC MONDAY UNDER 1600 1 win and 1 loss and then withdrew
1993
1-2nd U1400 Southern Open $?
World Open U1600 228th place
U1400 2-5th 20TH ANNUAL LIPKIN/PFEFFERKORN $?? ($400 would have been amount in 2003)
U.S. Open Thursday action quad G/10 25th out of 28
U.S. Open Friday G/1 Quad 1st-2nd place 3-0 $?
Atlantic Open U1400 16th place
U1400 US Quick Open Championship 38th place
1995
New York Open U1600 1 loss and 1 win and withdrew
Chicago Open U1600 1 win and 2 losses and then withdrew
World Open U1600 59th place 1 win and 2 losses and then withdrew
NEW YORK OPEN U1600 2nd-5th $1,000 approx
1996
World open U1800 1 loss and 1 win and then withdrew
1998
World Open U1800 1-2 place $7,500
2003
35TH LIBERTY BELL OPEN!!! U2000 1 win and 2 losses and then withdrew
12TH EASTERN CLASS CHAMPIONSHI U2000 1st place 4.5/5 $1,000
U2100 GAME/90 CHAMPIONSHIP MA 2 losses (1 to 1800 and 1 1200) and then withdrew
FOXWOODS U2000 3 losses and then withdrew
WORLD OPEN 8th-12th 3 buys, 1 loss and then 5 wins 6.5/9 $340
NORTHEAST CHESS FALL GETAWAY 3rd through 6th U2100 4/5 $1,300
NATIONAL CHESS CONGRESS 6.5/7 1st place U2200 5.5/6 $2,000
2004
FOXWOODS 4.5/6 21st place
WORLD OPEN U2200 2/5 and then withdrew
NATIONAL CHESS CONGRESS U2200 4.5/6 2nd-5th place $500 Approx
NORTH AMERICAN OPEN Tie for 3rd-9th $1,200 approx
2005
HB global 1 loss to 1600 player 1 loss to expert withdrew
North American Open Section 1 win and 2 losses and withdrew
NY JANUARY UNDER 2300 played one game and lost to U2000 player
FOXWOODS 1st place U2200 6.5/7 $4,015
2006
National Open 40th place in open section rating 2200 3.5/5 but uscf 2144
World Open section 97th place
I still don’t see where $100,000 is coming from…
Anyway, I believe the big CCA (Bill Goichberg) tournaments now have a clause in them that states if you have been above the rating class for the section you are currently playing in within the last 12 months (e.g. if you’re in the U1800 but your rating 10 months ago was above 1800) then you can only win $1,000 maximum and the remainder of your prize gets shared amongst the other players involved.
Chris Bird