The Imaginary World of Ratings

According to USATT, http://www.usatt.net/league/singles/ratings_league.html if you are rated below 1000, you are a beginner. Your results may vary. I once played a 500 rated player who was rerated to 1300 following the tournament. I know several under 1000 players who if you only watched them warm up, you would assume they were rated at least 1500. That doesn’t mean they can play at a 1500 level, but they don’t really look like beginners either.

 

Under 1000 Beginners
1000-1700 Intermediate
1400-1700 Average member of USA Table Tennis
1700-1800 Average tournament player
2000 An “Expert” player
2200 A “Master” player
2350-2650 Members of the USA Women’s National Team (5 players)
2550-2800 Members of the USA Men’s National Team (5 players)
2800-3000 The best players in the world

 

Incredibly, if you can get one point over a 1000 rating, you can consider yourself an intermediate player. Of course you may very well be intermediate for quite some time. In fact, should you fight your way to a 1400 rating, you can now consider yourself average. Actually, you aren’t even an average tournament player. At 1400 you are on the low end of average USATT members.
Now, should you make it to a 1700 rating, USATT will finally consider you truly average. You may beat the typical person on the street, and all but five of your local club members, but among tournament players, you’re strictly middle of the pack. A 1800 rating is still average. I currently have a 1790 rating. USATT used to consider me above average when I was rated 1855, but sadly, I am back to being average after a year of 1800 bliss. I still have a club rating over 1800 at AGTTA, so if I want to feel better than average, I can always go there.
If you happen to be rated between 1800 and 2000, there’s still some bad news. The USATT ratings chart has no real category for you. If you were over 2000, you could be considered an expert. Until then, you are left in the 200 point limbo of being extremely average. I recently played an expert, who had been an expert since 2012, when he played his last tournament. Knowing he’s rated 2001 does seem to make any wins I have over him sweeter. It does occur to me that if he played a tournament and lost two rating points, would he really not be an expert anymore?
For many players, reaching a 2000 rating is the ultimate goal. Sure, if you made it to 2200 you could be a master, and a 2800 rating might allow you to make a living playing tournaments. 2000 is a nice round number and probably insures enough ability to beat all the average players. Right now, I’d like to get to 1900. On my best days I imagine I’m playing at least at a 1900 level. Had I known it was possible to get a 1900 estimated rating in my first tournament, I could be 1900 right now, without ever winning a match. I actually witnessed this at a recent tournament. The player with the 1900 initial rating was far closer to 900.
The rating system we use is not perfect and sometimes seems counterproductive. While it is supposed to be an objective measurement of players abilities, sometimes it doesn’t work out that way. Not that I have any better system in mind, I’d love to have a higher rating. It is helpful to have some perspective on ratings and what they actually reflect. We can be unique and special players and people. Let the ratings fall where they will.

3 Replies to “The Imaginary World of Ratings”

  1. “While it is supposed to be an objective measurement of players abilities, sometimes it doesn’t work out that way.” The problem is not just whether it is objective, but whether it is accurate–or more to the point, how accurate it is.

    It is important to understand what the rating means. It is an estimate of your ‘true’ rating (the rating your would have if the system were perfect). This estimate comes with an error, ‘the error of estimate’–which is unknown ( but there are ways to estimate that error). It’s better to think of the rating as having a confidence interval around it. So a rating of, say, 1500, means that while 1500 is our best guess of your ‘true’ rating, that ‘true’ rating lays somewhere between, say, 1300 and 1700. Of course, your ‘true’ rating is more likely to be between 1400 and 1600, than it is between 1300 to 1400 or between 1600 to 1700. But there is also a small probability that your true rating is below 1300, or above 1700. And of course, the error of estimate is greater depending on the reliability of the data used to estimate the rating. In other words, the rating of a regular competitor, based on a lot of competition, is likely to be more dependable than the rating of an occasional competitor, who has only played a small number of people.

    1. Hi Gary. Nice explanation of the rating system. You are, of course exactly right. I just sometimes find the outliers interesting. When you consider how much stock players put in ratings, it was an easy target for a satirical look at them.

      1. Hi Jon, Yes, even without people trying to ‘game’ the system, there are many reasons to expect the ratings to be unreliable, especially for infrequent competitors, whose ratings will be based on a small number of matches that may not be a fair representation of their current ability. It does not make sense to put too much confidence in them. But I gather also that there are many cases of people trying to ‘game’ the system, in order to keep their ratings unnaturally low: perhaps to win a targeted competition, or to protect young competitors from being beaten too often by better players. In these cases, the ratings are not just unreliable, but essentially meaningless.

        If you are a regular competitor, or are coaching competitors, no doubt studying the outliers is very useful. Given that players probably put far too much stock in these ratings, pocking fun at them seems like a useful service to the community.

Leave a Reply

Your email address will not be published. Required fields are marked *