Moon Shooter
Hall of Fame
FYB
I played 5 mixed doubles matches in October and November and my UTR will vary from 3.3X to a high of 6.0X since I last played.
I think Max G. Identified the biggest problem - the 12 month cut off. Keep in mind that this is not only a problem for the person directly but this cut off means the ratings of everyone you played or partnered with is constantly changing as well as everyone they played with or against etc. Again the importance of earlier matches that were played over 12 months ago could fade to zero if someone has many current matches. But if they don't have any or very few matches in the last 12 months using the older matches is obviously going to give you better results. If you said "I have no idea how good this player is" when you knew they had a 11.63 UTR 13 months ago you would be called stupid (or worse dishonest). Of course, that is valuable information that UTR chooses to ignore.
The other problem (which compounds the first problem) is the complete divide between singles and doubles play. This is great if someone has lots of both types of play in a 12 month period. But if someone has only 1 or 2 in one category and a whole bunch in the other it should factor in. If you have a 6.34 based on 30 doubles matches you should not have a singles rating of 3.74 based on 2 matches. The algorithm should allow Singles ratings to be effected by doubles ratings if you have below a certain number of singles matches and many doubles and vice versa. This effect can fade out to nothing as the number of matches in the particular category increases.
Finally as long as the rating system has a cap and bottom the rating system will either need to sacrifice accuracy or it will need to bounce around even when you are not playing. I'm not sure how much of an effect this would have overall (probably not much) but I also don't see any upside of having a cap and a floor in ratings. The floor and cap have other downsides besides meaning the ratings should constantly bounce around even if you are not playing.
If they made the algorithm address these issues I bet my rating would only vary about .5 points instead of the 2.7 points we see now. The entire system would also be more accurate - that is better able to predict the score of future rec level matches.
In your area people likely have many matches that are put into UTR. And for you UTR will work well as it is. But in areas like mine where likely less than 5% of the matches played are recorded in UTR the system will not work so well for most players. Because it does not work well, people are not interested in it and therefore people don't bother entering matches. It is a cycle that my area needs to break but it would be easier to break the cycle if UTR made some sensible changes.
I played 5 mixed doubles matches in October and November and my UTR will vary from 3.3X to a high of 6.0X since I last played.
I think Max G. Identified the biggest problem - the 12 month cut off. Keep in mind that this is not only a problem for the person directly but this cut off means the ratings of everyone you played or partnered with is constantly changing as well as everyone they played with or against etc. Again the importance of earlier matches that were played over 12 months ago could fade to zero if someone has many current matches. But if they don't have any or very few matches in the last 12 months using the older matches is obviously going to give you better results. If you said "I have no idea how good this player is" when you knew they had a 11.63 UTR 13 months ago you would be called stupid (or worse dishonest). Of course, that is valuable information that UTR chooses to ignore.
The other problem (which compounds the first problem) is the complete divide between singles and doubles play. This is great if someone has lots of both types of play in a 12 month period. But if someone has only 1 or 2 in one category and a whole bunch in the other it should factor in. If you have a 6.34 based on 30 doubles matches you should not have a singles rating of 3.74 based on 2 matches. The algorithm should allow Singles ratings to be effected by doubles ratings if you have below a certain number of singles matches and many doubles and vice versa. This effect can fade out to nothing as the number of matches in the particular category increases.
Finally as long as the rating system has a cap and bottom the rating system will either need to sacrifice accuracy or it will need to bounce around even when you are not playing. I'm not sure how much of an effect this would have overall (probably not much) but I also don't see any upside of having a cap and a floor in ratings. The floor and cap have other downsides besides meaning the ratings should constantly bounce around even if you are not playing.
If they made the algorithm address these issues I bet my rating would only vary about .5 points instead of the 2.7 points we see now. The entire system would also be more accurate - that is better able to predict the score of future rec level matches.
In your area people likely have many matches that are put into UTR. And for you UTR will work well as it is. But in areas like mine where likely less than 5% of the matches played are recorded in UTR the system will not work so well for most players. Because it does not work well, people are not interested in it and therefore people don't bother entering matches. It is a cycle that my area needs to break but it would be easier to break the cycle if UTR made some sensible changes.
