An alternative rating system to the existing USATT system, Marcus, and others, based on total points per match instead of winning or losing the match, has the following advantages:
Easy to calculate without software, or even a calculator, and intuitive. Rating points have a real world meaning, i.e. opponents’ ratings predict how likely a player is to take a point. A player that is rated, for example, double his opponent, is expected to win twice as many points in the match.
It is easy to calculate a handicap, with no chart necessary. If the higher player is rated, say, 34, and the lower player rated 32, and they are playing a 51 point handicap match, the lower player gets 3 points. The obvious reason is that the higher rated player is expected to take 34/32 (or 51/48) of the match points.
The system can be used to rate in individual’s performance in doubles matches, or even a combined singles / doubles rating for a player. This works because the comparison between two players is a ratio, so if, for example, a player rated 50 plays with a partner rated 40, the higher player is presumed to be responsible for 5/9 of the post match rating, and receives 5/9 of the rating points.
The system works as follows:
The average player would be rated at approximately 50. The lowest possible rating is 0. There is no theoretical limit on the high end of the rating scale, although it is anticipated that triple digit ratings would be rare (to prevent overall rating deflation, it would be wise to use a Canadian style 20 match temporary rating, i.e. the first 20 matches a new player plays don’t affect his opponent’s rating).
It is critical, for the system to work, that players keep an accurate record of the points scored in each game. This is done for tiebreaking purposes in any event, and it is probable that players would be far more meticulous in keeping an accurate record if they know that the rating is calculated based on match points.
The calculation for an individual match consists of adding the opponents’ ratings, and then awarding each player the same ratio of the available rating points as each scored match points. For example, if two opponents are both rated 60, and one scores 33 points in the match and the other 27 points, the winner would, at the end of the match, be rated 66 and the loser 54. It should be noted, however, that if a heavily favored player beats an underdog by a small margin, the favorite can lose points, and the underdog gain, even though the favorite won the match. In the above example, if the favorite starts the match with a 70 rating and the underdog has only a 50 rating, and yet the favorite only scores 33 points to the underdog’s 27, then the ending rating is still 66 for the winner and 54 for the loser. Therefore a close match would be rated differently than a lopsided match, despite the same match result.
Ratings based on doubles matches can be calculated the same way, with one additional step alluded to above. For example, doubles partners rated 60 and 40 play another team rated 55 and 50. The total rating points to be divided in the match would be the sum total of all the players (205 in this example). If the first team scores, say, 42 match points and the second team 38 (e.g. 11-9, 9-11, 11-9, 11-9), then the rating points to be divided between the members of the first team is 205 x 42/80= 108, and the rating points to be divided between the members of the second team is 205 x 38/80 = 97. The first team would divide its 108 points by splitting them on a 60/40 basis between the stronger and weaker player, i.e. 65 points would be the new rating for the stronger player, and the new rating total for the weaker player would be 43 points. Similarly calculated, the losing team’s new ratings would be 51 and 46.
Assuming, as is usually the case, that a player plays multiple matches at a single tournament, all one would need to do to calculate that player’s rating is to calculate the mean of the ratings emanating from each match.
Published ratings after a tournament could include a singles rating, a doubles rating, and a combined rating. One could even include hardbat or sandpaper ratings in the combined rating, if desired.
/ Cliff Falk
|