I have seen tournaments advertising that they will use the higher of people’s quickplay and regular ratings for an event, presumably to stop someone playing in a lower section than they really should be in. I have a few questions in relation to this:
Is there a way to get WinTD to do this automatically? I know WinTD can use someone’s regular rating if they don’t have a quick-chess rating, but not how to get it to use the higher of the two. Also, is there a way to get it to use the quick-chess rating if they haven’t played a regular rated game?
I know of some players (mainly older) who are pretty good at regular chess, but play very badly or lose a lot of games on time in quick-chess. Why would I want to use their regular rating for a quick-chess tournament like this when they have a lower rating in that category because they really are a worse player in shorter time controls? (A point that is going to be even more so highlighted since blitz chess also falls into the quick-chess category.)
I’m not sure I agree with this idea, but the reasoning behind it is clear. Most people have played very few Quick Chess games (sometimes years ago), and the Quick Chess ratings for such players are highly inaccurate. (Before someone starts arguing with me about this, I readily agree that there are local exceptions to this rule. It is true in most areas, however.)
The USCF’s about-to-be-retired ratings data structure had no direct way to track dual-rated events. (The new ratings system does.)
However, I did some analysis on it about 18 months ago and concluded that only about 10% of all of the quick rated games were quick-rated-only, the rest were dual-rated.
That percentage may have gone up a bit when blitz events became quick ratable, but I doubt if it’s up beyond 25%.
We have about 29,000 players with established regular ratings that have been updated in the last 12 months. We have about 12,750 with recently active established quick ratings. About 11,800 of those have both an active established regular rating and quick rating.
Yes. The “Section setup” panel has an option to “Use higher of main or local rating.” By “local,” they mean “alternate” or “secondary,” which in the USCF database is the Quick rating by default. I suspect the original idea was that users would configure their own database with club ratings, but I doubt many do so.
I don’t know that using the higher of the two ratings makes much sense unless you assume they’re measuring the same skills, which clearly they are not.
If both ratings are fairly current, they’re going to be similar, though there is a tendency for the regular rating to be higher than the quick rating.
Of those who have both a fairly current established regular and quick rating (figures given yesterday), for 8119 of them the two ratings are within 100 points of each other.
259 have a quick rating 100-199 points higher than their regular rating and another 50 have quick ratings 200 or more points higher than their regular rating.
For 2324 of them the regular rating is 100-199 points higher than the quick rating, for 753 the regular rating is 200-299 points higher than the quick rating, and for another 300 the regular rating is 300 or more points higher than their quick rating.
I’m not supporting this policy; in fact, I don’t use it. However, those who do can point to a number of cases with gross differences between the ratings. Most of these are people who played one or two Quick Chess tournaments years ago. Allowing these people to play with “Quick” ratings can cause real distortions in the prize distribution.
I agree that the difference is getting smaller with “dual rating,” but there are two things to bear in mind here. First, this does not necessarily represent any increase in the popularity of “Quick Chess.” Secondly, if the two ratings measure different skills (widely assumed but not proven), then then adding to the Quick rating pool with dual-rated G/40 tournaments will not improve the accuracy of the Quick ratings.
Perhaps the slight inaccuracy of including G/30 to G/60 events (due to the possibly different skills being measured) is offset by the greater size of the data set on which to base the computations.
I don’t view it as strictly an anti-sandbagging measure. Although I don’t do this, I’ve done something similar … In a club that’s growing quickly, your first bunch of tournaments are going to have a LOT of unrated players. Since the Swiss system works “best” when the pairings are in order by strength (or so my deluded thinking goes!) what do I do with all the unrateds I have from my club that attend their first tournament? Well … I think I’m allowed to assign ratings to them to approximate their strength … but that can appear arbitrary. If they have a quick rating and NOT a regular rating I usually use the quick rating for them.
Once I’ve committed this sin (?) it’s a small step to conduct a tournament where you use the higher of the ratings. Especially if it’s going to be dual rated and then you have all the philisophical justification you need. In fact - following this argument - I’m surprised no one has averaged the two ratings together to determine pairing numbers!