Moon Shooter
Hall of Fame
This is intended as a discussion where we can talk about all the pros and cons of how USTA uses its rating system and what some solutions might be. I started it as a response in a thread but it is clear this topic is much broader and I don’t want to hijack the other persons thread more than I have already.
First I would like to thank Schmke for his latest charitable comments. I was afraid that we were losing patience with each other. I will try to do the same.
Lets me start with I think an important point he made:
I think this is important because there are many different problems and how potential solutions may or may not effect these different problems will vary depending on what the problem is.
Here are the problems I see with the how USTA uses it's rating system:
1) Exclusion: It rigs the level based national championship so the majority of tennis players have no real chance to win. This is because it favors very densely packed tennis communities that will happen to have several players right below the bubble cut off. This makes it difficult for smaller communities to even form teams.
2) The rating system is widely considered bunk. No one knows what it even is supposed to mean to be a 4.0 so no one cares.
3) It gives the appearance of corruption. They allow some appeals and deny others and no one knows why.
4) The only actual use for the rating system is to disqualify people from playing with their friends so it greatly encourages sandbagging. For example if you just made it into the 4.5 level congratulations no one will want you on their team! Why? Because you are likely dynamically rated about 4.0X and so you no longer qualify for any 4.0 teams and no 4.5 team will want you because you will lose to all the other 4.5 players who are 4.4X. So if you want to be on a team without killing their chances at nationals your only options are to appeal or sandbag.
I will refer to these problems in my post below. Notice I do not claim that the USTA algorithm is faulty. The actual algorithm that they use to get the dynamic ratings seems ok. It is rather how they use this algorithm that is the problem.
My proposed solution to these 4 problems is to have USTA publish the full dynamic ratings and take the average of the players that actually play for a team instead of requiring every single player to be below a certain line. The next post is why I think this will greatly reduce these problems and also address concerns that this solution would create other problems.
First I would like to thank Schmke for his latest charitable comments. I was afraid that we were losing patience with each other. I will try to do the same.
Lets me start with I think an important point he made:
We should perhaps clarify what "these issues" are.
I think this is important because there are many different problems and how potential solutions may or may not effect these different problems will vary depending on what the problem is.
Here are the problems I see with the how USTA uses it's rating system:
1) Exclusion: It rigs the level based national championship so the majority of tennis players have no real chance to win. This is because it favors very densely packed tennis communities that will happen to have several players right below the bubble cut off. This makes it difficult for smaller communities to even form teams.
2) The rating system is widely considered bunk. No one knows what it even is supposed to mean to be a 4.0 so no one cares.
3) It gives the appearance of corruption. They allow some appeals and deny others and no one knows why.
4) The only actual use for the rating system is to disqualify people from playing with their friends so it greatly encourages sandbagging. For example if you just made it into the 4.5 level congratulations no one will want you on their team! Why? Because you are likely dynamically rated about 4.0X and so you no longer qualify for any 4.0 teams and no 4.5 team will want you because you will lose to all the other 4.5 players who are 4.4X. So if you want to be on a team without killing their chances at nationals your only options are to appeal or sandbag.
I will refer to these problems in my post below. Notice I do not claim that the USTA algorithm is faulty. The actual algorithm that they use to get the dynamic ratings seems ok. It is rather how they use this algorithm that is the problem.
My proposed solution to these 4 problems is to have USTA publish the full dynamic ratings and take the average of the players that actually play for a team instead of requiring every single player to be below a certain line. The next post is why I think this will greatly reduce these problems and also address concerns that this solution would create other problems.
Last edited: