My US Chess rating experiment

All Things Chess
1

US Chess has always had the once rated always rated policy. And I don’t remember where but someone had once said that any errors in rating would work there way out within I think 25 games or so. But anyway I if nothing else but for curiosity I took my last 24 game results and used the US Chess rating estimator to see what my rating would be now if I had only began play 24 games ago. I did it in two batches of 12.

I entered the first batch and had 8 wins and 4 losses. Doing my normal rating and leaving it at 50 games played my rating would have come out to 1596. Unrated it would have been 1601. I then used that 1601 number and entered the second batch of games with a score of 8.5. Here I put the number of previous games at 12. This gave me a rating of 1713 after 24 games. My actual current rating is 1678. That seems pretty close to me.

So I would imagine that using the once rated always rated thing doesn’t really matter all that much in the long run, either way.

But then that may just be me. :slight_smile:

2

I agree, in the long run, but it may influence class/under prizes tremendously in the short run.

Alex Relyea

3

Maybe in a good way in certain circumstances. Say player played 10 years ago as a kid with a 600 rating. Hasn’t played since then but now is getting back into it. But his or her real strength is now 1600. As a 600 rated player he or she would be eligible for prizes that he or she wouldn’t be as an unrated. Not only that when he or she wins against 1500 or so rated players they lose rating points that they wouldn’t have if they had started out as unrated.

Or look at the other side. Older player 1800 reentering after years of no play. Playing now at 1600 level. Still as unrated uneligible for class or under prizes. Eventually ends up where he should be after provisional rating.

4

You confirmed that statement in practice with your example, but it’s also true if you look at it in theory.

It’s easiest to see if you use the old 16 plus-or-minus 4% linear approximation, which corresponds roughly to a K-factor of 32. Under that approximation, a 1500 who draws a 1700 will gain 8 rating points, because 8 is 4% of the 200-point rating difference. (The “16 plus-or-minus” part doesn’t enter the picture if the game is drawn.)

Suppose the 1500 enters a 25-round tournament and gets 25 draws, all against opponents rated 1700. He will gain 8 points per game, for a total gain of 8 times 25, which is 200 points. His post-tournament rating will be 1700.

Meanwhile, his performance rating (i.e. his rating based on just the one event) will also come out 1700. (What else? He has just drawn 25 players, all rated 1700.)

So, with a 25-round tournament, the player’s rating has caught up completely with his performance rating. The 200-point “error” has worked its way out within 25 games, just as you said above.

It follows that, with a 30-round tournament for example, the player’s new rating would actually overshoot his performance rating. The player will end up at 1740, forty points higher than he “deserves”. That’s one reason (I suppose) that the US Chess rating software limits the number of rounds per tournament. (I think the limit is 14 or 20 or something like that.)

Going back to the 25-round example, if the tournament is broken up into five 5-round events, then the correction will come more slowly, because the post-event rating for each 5-round event becomes the pre-event rating for the next event. The player will jump from 1500 to 1540 in the first event, but it will not go to 1580 in the second, because the rating difference (between the player and his five 1700 opponents) is now only 160 instead of 200. 4% of 160 is 6.4 (instead of 8.0), times 5 rounds is 32 (instead of 40), so after his second event he will be at 1572 instead of 1580.

Because of this smoothing action, the number of games required to completely wipe out any error will typically be a little more than 25 – perhaps around 35 or so.

Bill Smythe

5

I think Prof. Glickman has said that in theory it takes about 50 games for an anomalous or erroneous event (eg, a tournament rated under the wrong ID) to be fully attenuated out. I’ve looked at a number of examples of ID corrections entered well after the tournament ended, and it seems to me that in practice it happens faster, perhaps as fast as 25 games.

Also, keep in mind that for a mid-level player (1600-2000) the accuracy of a rating is probably +/- 50 points.

6

Probably because many players have K-factors larger than 32, and/or because of the two-pass system, and/or because bonus points are often involved.

Bill Smythe

7

Bonus points may be involved during the attenuation process if the anomalous event results in a noticeably lower rating for a player, but not if it results in a noticeably higher rating.

Some of the systems that have been suggested for dealing with stale ratings could, in effect, result in ‘negative bonus points’ to a player whose recent results suggest that the stale rating is significantly too high.

8

Has a rating ever been reset to zero for unusual circumstances like after a traumatic brain injury? I thought someone at least posted that question years ago.

Otherwise, I can’t really see a reason to reset a person’s rating. A single large monetary win at a tournament will give a person a ratings floor regardless of the person’s original rating. Not so unusual with kids. Say a kid played a couple years, then didn’t play rated USCF games for years, but rather just playing online for years, then in college win at some large OTB tournament prize… it would be expected that the kid playing unrated games for 10 to 12 years would have picked up some serious chess skills in the interim.

I kept revising the message, so I hope it looks ok. I have a broken arm at the moment and got tired of hunt-and-peck typing with one hand.

9

I think there have been a few cases where someone’s floor was lowered by 200 or 300 points retroactively, for a variety of reasons, but I don’t recall any cases where some’s rating was set back to 0/0 except if that was to correct for games rated under the wrong ID.

10

Ahh, thanks.