My .02 on Rating Pools and Stale Dates

[Steve Douglas]

My .02 on Rating Pools and Youth Chess

There is a discussion going on at the virtual AGM on the CFC discussion board regarding a potential problem in the rating system. The matter being discussed seems to relate directly to something I have observed and experienced myself and which I know has been an issue with many adult players I've talked to. I thought I would put my .02 in here in case it may be helpful in the discussion.

One of the things that all players pay too much attention to is their rating. Despite its being a measure of "relative" playing strength, most players tend to treat it as a measure of "absolute" playing strength. Players are unhappy if their rating goes down. One player I knew several years ago was upset when he won one of the club's tournaments with a score of 6/8 but his rating dipped slightly. He said he thought it wasn't allowed for your rating to go down if he won the tournament.

The problem seen, at least in the past, by many adult club players is that of "rapidly improving juniors". Once upon a time the *only* rated chess games you could play in Canada (outside Quebec) were CFC rated. Now, for juniors we have CMA events and of course the internet. Thus you can get the situation that a junior can go months or years playing competitive chess, improving the entire time, while their CFC rating stays static.

Another compounding factor is that with "junior-only" events, even if they are CFC-rated, you're still dealing with a closed pool of juniors. If they all improve at the same rate as a group, their ratings will not rise as fast as their playing ability. Thus when they "graduate" to "adult" play, their CFC ratings may be out-of-whack with their playing strength within the larger pool of all players in Canada.

This can become a problem, or a source of resentment, in some circumstances. I will give a couple of examples from my own experience (going just from memory).

Years ago I took my son to play at a local chess club. It was a small club with approximately 20 members and about 14 of those would be present in any given week. My son already had a provisional CFC rating of about 1100 based on 5 games or so. Only a few members of the club played outside the club. It was a pretty small pool. Also, almost all the members were adults, and as we all know, once a chess player gets past a certain age, playing strength isn't going to improve much unless they work really hard at it. After about 3 months at the club, my son had a regular rating. During this period he didn't siphon too many rating points from the club pool of players. He played in some tournaments and the provisional rating system did its job. His rating quickly adapted to his playing strength.

Now for another example. A few years later we had a college-age student join the club. His CFC rating showed him to be a class C player. He finished second in the tournament beating everybody but an expert-level player. When I looked up his rating, it was still a provisional rating with about 18 games played, with all of the games coming from Ontario High School Championship tournaments played over several years with the most recent being two years previous. In short, when he finally got his permanent rating, it was based mostly on games that were many years in the past. Despite showing Class A or even Expert strength, he got a newly minted permanent rating showing Class C. He siphoned a lot of rating points from the club in the next period because the bonus-point system wasn't good enough to bring his rating up to his playing strength quickly enough.

The second example caused a lot of grumbling amongst some of the club members. Nobody likes to see their rating drop like a stone.

The problem in this second example could be avoided if their were a way to "stale date" either ratings or games. If the older OHSCC games for this second player had been tossed out, his first 24 games at the club would all be under the provisional system and there would not have been a siphoning of rating points and a depression of ratings in that (admittedly small) pool. I hope some of the governors/voting members who are involved in that discussion at the AGM find this info helpful.

(I have some opinions regarding CFC/CMA imbalances but I'm already being more long-winded than usual.)

Steve

[Paul Leblanc]

Re: My .02 on Rating Pools and Stale Dates

Steve, you have put your finger on the single biggest problem I face.

We are dealing with some legacy issues from the days when juniors earned their 3-digit ratings by playing lots of 30 minute games against each other then showing up at open events and playing 400 points above their rating.

The bonus point system has fixed many of these anomalies. I'm hoping that we can come to agreement on another initiative to re-set CFC ratings of Canadian players who play in USCF or FIDE events and achieve a rating far above their CFC rating.

Going back and erasing some past results is a bit too bold for me but it sure is tempting!

[Steve Douglas]

Re: My .02 on Rating Pools and Stale Dates

Steve, you have put your finger on the single biggest problem I face.

We are dealing with some legacy issues from the days when juniors earned their 3-digit ratings by playing lots of 30 minute games against each other then showing up at open events and playing 400 points above their rating.

The bonus point system has fixed many of these anomalies. I'm hoping that we can come to agreement on another initiative to re-set CFC ratings of Canadian players who play in USCF or FIDE events and achieve a rating far above their CFC rating.

Going back and erasing some past results is a bit too bold for me but it sure is tempting!

Hi Paul:

When I get a chance and am not busy pontificating on political nonsense, I'll go back and find the specific cross-tables for the events. The main thing I'm proposing is stale-dating games that contribute to provisional ratings. If a player goes through high school and plays in 4 OHSCC's, they will likely play 23 games, not enough for a permanent rating. Their performance rating for each tournament will likely be: 1100, 1300, 1450, 1600. And since they maybe playing against *other* players who only play in the OHSCC, then the ratings may be even more meaningless. In the example above, I would likely only include the last 11 games (1450 and 1600) and keep a "rolling" provisional rating until you get a cluster of games played close together. After that the bonus system should work.

Steve

[John Brown]

Re: My .02 on Rating Pools and Stale Dates

Why not just start any player with a 3 digit rating (who wants to play in a CFC rated tournament )with a rating of 1400.
That will be their new CFC Rating.
If they do well their rating goes up if they do poor their rating goes down. It will have a bottoming out level that it can never go back to a 3 digit rating again. The bottom rating could be say 1150.
This way if anyone who has a rating under 1600 will not lose mega points every time they play these 3 digit rated players.
You want players to play chess. They won't play if their ratings get abused.

[Paul Leblanc]

Re: My .02 on Rating Pools and Stale Dates

Steve, I'm certainly willing to have a look at the idea

[Vlad Drkulec]

Re: My .02 on Rating Pools and Stale Dates

Why not just start any player with a 3 digit rating (who wants to play in a CFC rated tournament )with a rating of 1400.
That will be their new CFC Rating.
If they do well their rating goes up if they do poor their rating goes down. It will have a bottoming out level that it can never go back to a 3 digit rating again. The bottom rating could be say 1150.
This way if anyone who has a rating under 1600 will not lose mega points every time they play these 3 digit rated players.
You want players to play chess. They won't play if their ratings get abused.

That is an interesting idea.

[Bob Armstrong]

3 digit players - at Scarborough CC

Just to provide some facts on this from Scarborough CC. We play in four sections (over 100 players per regular Thursday night Swiss). The bottom section is U 1400.

For the last tournament of our 2013-4 season, there were 51 players in that section (almost half the club; and likely 80% of them were juniors).

Of the 51 players, 65% (33/51) were UNDER 1150.

So, in our set-up, the three digit players play most of their 6 tournament games, against other 3 digit juniors. The highest rated player the 3 digit player gets to play is 1299 (because the 1300's all play up in the U 1600 section - they have an option). And this is likely only in the first round. When they lose, they end up playing much lower rated players for the last 5 rounds.

So we have some 3 digit players upsetting players over 1150, but only in the bottom section. But it does prove hard on the 1150-1299 top players in the section. There are all these up and coming 3 digit players nipping at their heels.

Perhaps Paul may be able to use these facts as part of his evaluation of the proposal.

Bob A (SCC FB Page Co-Manager)

[Gary Ruben]

Re: 3 digit players - at Scarborough CC

The ELO rating system is the worst thing that ever happened to chess.

It seems players attain the rating they want, or fail to attain what they think they should and they quit. Few seem to play for fun anymore.

Playing in rating sections is hideous. Players simply play in rating ghettos and a TD doesn't have the ability to try to develop players by pairing them against stronger opponents

If someone has a different opinion, I'm not interested in hearing it. :D:D

P.S. I find the computer detector police beyond the pale.

[Nigel Hanrahan]

all hail the ELO system.

The ELO rating system is the worst thing that ever happened to chess.

It's a means of comparing a player's strength to someone else's playing strength. An objective criteria. What's wrong with that?

It seems players attain the rating they want, or fail to attain what they think they should and they quit. Few seem to play for fun anymore.

Most chess players never play competitively. They are the overwhelming majority of players. Any competitive sport is like that; players only play in competition when they feel prepared. Especially if they know what it takes to win competitively.

Playing in rating sections is hideous. Players simply play in rating ghettos and a TD doesn't have the ability to try to develop players by pairing them against stronger opponents.

Playing in rating sections is fabulous. By playing against roughly equal players (or slightly stronger), a player gets a chance to improve by playing against their "peers" in chess. Nothing is more wasteful than the tiresome weekend swiss in which an aspiriing player alternately plays mismatches, first against someone much stronger than them, then someone much weaker than them, and so on. Now THAT'S hideous.

If someone has a different opinion, I'm not interested in hearing it. :D:D

Sorry. I don't like to read your contributions more than once. I didn't read this bit until it was too late.

[Gary Ruben]

Re: all hail the ELO system.

It's a means of comparing a player's strength to someone else's playing strength. An objective criteria. What's wrong with that?

You don't need ratings to determine strength. Whoever leaves with the most prize money or the trophy is the best player that day.

Most chess players never play competitively. They are the overwhelming majority of players. Any competitive sport is like that; players only play in competition when they feel prepared. Especially if they know what it takes to win competitively.

If they play when they feel they can win competitively, why hold them back with stupid rating sections? If they don't have some arbitrary rating they can only play for best of the patsers.

Playing in rating sections is fabulous. By playing against roughly equal players (or slightly stronger), a player gets a chance to improve by playing against their "peers" in chess. Nothing is more wasteful than the tiresome weekend swiss in which an aspiriing player alternately plays mismatches, first against someone much stronger than them, then someone much weaker than them, and so on. Now THAT'S hideous.

Could that be because you're old and looking for a break? You mentioned knowing what it takes to win competitively then you dismiss part of the process. Beating stronger players and losing to weaker players because you don't know how to beat junk openings and traps is part of the game.

Sorry. I don't like to read your contributions more than once. I didn't read this bit until it was too late.

That part I wrote was about comments was a joke. I notice you've been on the defense recently for using the term idi.. Aw, shucks. ;)

[Fred McKim]

Re: My .02 on Rating Pools and Stale Dates

Why not just start any player with a 3 digit rating (who wants to play in a CFC rated tournament )with a rating of 1400.
That will be their new CFC Rating.
If they do well their rating goes up if they do poor their rating goes down. It will have a bottoming out level that it can never go back to a 3 digit rating again. The bottom rating could be say 1150.
This way if anyone who has a rating under 1600 will not lose mega points every time they play these 3 digit rated players.
You want players to play chess. They won't play if their ratings get abused.

I suspect 1400 is too high. I find the majority of children who join the CFC to play in open tournaments are initially between 1100 and 1300 strength.

In an earlier version of the rating system any player rated below 1200 and achieving a Rp > Ro and <= 1200 were simply set to that rating. For players with Rp > 1200 their pre-tournament rating was adjusted to 1200. This allowed an easy ascent to 1200 not siphoning mega points from the average players.

[Steve Karpik]

Re: 3 digit players - at Scarborough CC

Gary I think you're maybe being a little harsh about the impact of ELO-type ratings. I actually enjoy seeing what my rating is doing. It sort of gives me a sense of whether I'm improving or getting worse. Lately I've been spending way too much time on my business and my rating shows it. My chess rating is plummeting. I'm not too happy about that but it gives me impetus to spend a little more time enjoying chess and hopefully seeing my play improve. So although I'm not happy about seeing my rating hit new lows, I'm not upset and I'm maybe inspired.

However, I do have some concerns over how we've been handling Swiss tournaments at the Scarborough Chess Club. We've got multiple sections (U1400, U1600, U1800, 1800 ). If a player is within 100 points of the next level they can play up. The problem with this way of doing things is that talented new players get trapped in the lower sections for too long. They can't gain rating points fast enough because they are stuck playing weak, low rated players (like me). Eventually they get enough rating points to make it into a section that is reasonable for their playing strength.

A better idea is to do what we did for our club championship. We accelerated the first two rounds to avoid situations where a master would be bored playing a low rated player (like me). It worked out pretty well. The truly good players with established ratings ended up getting paired with other good players. Weak players ended up battling it out with other weak players. And most importantly, the young stars with unrealistically low ratings worked their way up the ladder so they were playing moderate to high rated players. Yes, some of the good players took a hit for a few rating points but with bonus points and the ELO formula, the talented youngsters will quickly have a rating commensurate with their abilities.

So I would defend the ELO system and ratings in general. But I also recognize that sections (especially when you've got too many sections) can ghettoize young talented players and slow the assessment of proper ratings for these players.

At our AGM this year, I will be urging the SCC to go to a single section format with acceleration. (For those who don't know our club, we have about 100 players active every week and their ratings range from about 800 to 2200 or so. In a smaller tournament, my ideas may not make sense.)

So in the end I both agree and disagree with you.

[Mathieu Cloutier]

Re: all hail the ELO system.

Playing in rating sections is fabulous. By playing against roughly equal players (or slightly stronger), a player gets a chance to improve by playing against their "peers" in chess. Nothing is more wasteful than the tiresome weekend swiss in which an aspiriing player alternately plays mismatches, first against someone much stronger than them, then someone much weaker than them, and so on. Now THAT'S hideous.

Agreed with most of what you wrote there. 'Yo-yo' pairings are a pain and waste of time in large open tournaments with one section. It doesn't affect stronger players, as they tend to win and play against increasingly stiffer competition. But it can be a real pain for guys rated between 1600 and 2000.

The only thing I find questionable about sections is that it does tend to create 'getthos'. A guy rated 1750 can win the -1800 section two, maybe even three times without seeing his rating go up enough to move to the next section. You can try to play up in the -2000 section, but it can be hard to adjust on the fly (i.e. within 5 games!) to the higher level of play. Then you might just loose some ratings points and it's gonna be even harder to move up.

So sometimes, a 1 section tournament can be interesting for these guys. If you're doing good, you'll play against higher opposition, if you're not, it's OK, you won't spend 5 rounds or more being tortured by stronger players.

I actually think it's interesting to have different tournament formats. We have both at the Montreal chess club: round-robins with 6 players grouped by rating all year, and an open tournament at the end of the year.

[Gary Ruben]

Re: 3 digit players - at Scarborough CC

However, I do have some concerns over how we've been handling Swiss tournaments at the Scarborough Chess Club. We've got multiple sections (U1400, U1600, U1800, 1800 ). If a player is within 100 points of the next level they can play up. The problem with this way of doing things is that talented new players get trapped in the lower sections for too long. They can't gain rating points fast enough because they are stuck playing weak, low rated players (like me). Eventually they get enough rating points to make it into a section that is reasonable for their playing strength.

A better idea is to do what we did for our club championship. We accelerated the first two rounds to avoid situations where a master would be bored playing a low rated player (like me). It worked out pretty well. The truly good players with established ratings ended up getting paired with other good players. Weak players ended up battling it out with other weak players. And most importantly, the young stars with unrealistically low ratings worked their way up the ladder so they were playing moderate to high rated players. Yes, some of the good players took a hit for a few rating points but with bonus points and the ELO formula, the talented youngsters will quickly have a rating commensurate with their abilities.

Accelerated pairing does seem preferable as a lower rated player has a chance given enough rounds.

With a low elo rating a player who is close to a chess addict has a better chance of improving quickly than a casual player who only plays from time to time and is stuck in a low rating pool. These days with so much online play I wouldn't be surprised if a lot of players are getting better games online than they would at a club.

The young players tend to get good fast. When I was pairing CC events by the time a year passed a young player was no longer playing at the strength he had played and I bumped him up. Often to the same class as he CFC rating. Back then many players went back and forth between OTB and CC or played both.

I notice you mention the problems with being a weak player. Do you know strong players also face a dillema? What to do when health starts to fail or age set in and playing strength declines. You're left not playing up to your rating or what is expected of someone with your title. To retire or become "an opponent"?

As silly as it sounds, because of rating and title I can qualify for a GM norm CC event. While my rating is 25 points short, if it does not lower the category of the event I'd be in. Having a title helps.

For invitationals, I'm prime. A titled player who has presumably lost his skills due to age but has maintained a decent rating (2425) so the organizers home country nominees will have a really good chance of making a norm. What I call an "opponent".

When I played junior chess I didn't worry about ratings. There weren't any, as far as I recall back in the 1950's. How good a junior was depended on how a junior did against the other juniors and the adult players.

[Ed Seedhouse]

Re: My .02 on Rating Pools and Stale Dates

Why not just start any player with a 3 digit rating (who wants to play in a CFC rated tournament )with a rating of 1400.

It would be much simpler simply to treat these ratings as provisional. Since they are earned in what amounts to a separate rating pool they should be treated as provisional when they in effect enter a new pool. This will allow established players to avoid losing large numbers of rating points from playing these folks while bringing their ratings into sync with the main pool in fairly short order.

Simply assigning them a rating on the basis of no actual information seems entirely misguided to me.