How to keep ratings as intended throughout USTA

[raiden031]

I find the whole endless debates on this board over ratings (what to self-rate at and WHEN, etc.) to just be rediculous. Its so hard to to figure out what each person is rated because when playing USTA, they must rate low in order to get accepted onto a team. But if they play in club events or socials, they have to rate high to avoid getting stuck playing with people way below their level. Also they might have to overrate in order for potential playing partners to give them the time of day when responding to ads. The question is, how to keep the ratings to mean what they were intended to mean?

I guess the starting point would be inside USTA, where ratings are probably most skewed because there is most at stake (ie. Nationals). So the first question is whether the current ratings are what they were intended to be, on average?

We know that Nationals are probably not going anywhere, but what about the idea of human verifiers within Nationals to actually decide what the players should be rated, and then the computer takes over to traverse it down to all the various league players? I know there are verifiers already, but it seems like they have little purpose other than to point out that there are sandbaggers, and therefore the rating system might need some tweaking.

My biggest question is how does USTA prevent the ratings from permanently becoming skewed in one direction since only the computer (and appeals) seems to have the power to change ratings??

Or how about not allowing self-rated players to participate in Nationals?

 

[Topaz]

Or how about not allowing self-rated players to participate in Nationals?

That hardly seems fair...if someone has earned the right to play at Nationals they shouldn't be denied the chance just because it was their first year in leagues, and had to self-rate. Not everyone self-rates inaccurately. Plus, people do get better throughout the year, you need to take that into account if you are going to look at Nationals.

Why not take a closer look at districts or sectionals first?

 

[Cindysphinx]

I think the ratings in our area *are* quite skewed. Even the numbers show it. This spring, there were:

68 2.5s
400 3.0s
313 3.5s
137 4.0s
31 4.5s (with only 18 being actual 4.5s)

Shouldn't that distribution be a bit flatter? Shouldn't this be achieved by pushing some of the 3.0s and 3.5s up?

I also think it is weird that the ladies bunch up at 3.0 but the guys bunch up at 3.5 and 4.0.

Don't know if that is what you were driving at, Raiden, but there you have it.

 

[Topaz]

I think the ratings in our area *are* quite skewed. Even the numbers show it. This spring, there were:

68 2.5s
400 3.0s
313 3.5s
137 4.0s
31 4.5s (with only 18 being actual 4.5s)

Shouldn't that distribution be a bit flatter? Shouldn't this be achieved by pushing some of the 3.0s and 3.5s up?

I also think it is weird that the ladies bunch up at 3.0 but the guys bunch up at 3.5 and 4.0.

Don't know if that is what you were driving at, Raiden, but there you have it.

What about 5.0s Cindy? We've actually got some computer rated 5.0s, and occasionally they can field two teams and have a league!

 

[Raiden.Kaminari]

Or how about not allowing self-rated players to participate in Nationals?

There are pros and cons with this idea.

USTA wants to grow tennis, so they will still allow self-rated players to play at Nationals.

Plus, you would eliminate the 2.5 Nationals, since most true beginners would self-rate there. That brings up the question, why is there a 2.5 Nationals, anyway? Probably to help motivate goal oriented players, who want to improve. However, one common complaint is that at 2.5, most of the players are really 3.5 (where many of their year end ratings finish). A true 2.5 can be expected to improve to low 3.0 by the end of the season if they practice a lot. However, most adult 2.5 players rarely improve that much because of real life (who can spend so much time on the tennis courts?).

The best thing to do is to abolish Nationals until the rating issues are brought inline. Verifiers at Nationals would be nice as long as they have two verifiers from different sections to correlate their observations. The verifiers would have to observe different sections to prevent any conflict of interest. However, this would mean Nationals spending at least 34,000 per event (estimated cost of lodging/travel is 1000 per person, two verifiers from each section, excluding paying the verifiers for their time).

 

[Islandtennis]

I think the ratings in our area *are* quite skewed. Even the numbers show it. This spring, there were:

68 2.5s
400 3.0s
313 3.5s
137 4.0s
31 4.5s (with only 18 being actual 4.5s)

Shouldn't that distribution be a bit flatter? Shouldn't this be achieved by pushing some of the 3.0s and 3.5s up?

Cindy, I don't think you are skewed at all. In the Southern Section, the distribution is:

2.5 - 3%
3.0 - 33%
3.5 - 40%
4.0 - 19%
4.5+ - 8%

Your's seem to be in line.


Raiden, I agree with your idea of not letting self rated players play in Nationals. I would even go as far as saying that they shouldn't be allowed to play at District Championships. The biggest problem that I see with self rated players misrepresenting their rating is in teams that are trying to build a "super team" to play in Districts and above.

 

[JLyon]

Raiden, I agree with your idea of not letting self rated players play in Nationals. I would even go as far as saying that they shouldn't be allowed to play at District Championships. The biggest problem that I see with self rated players misrepresenting their rating is in teams that are trying to build a "super team" to play in Districts and above.

So basically you are giving new league players a big FU. Everyone has to start somewhere, if they were a previous league participant ok, I see the point but you will run off more than you keep by not allowing first year league players to advance with their teams.

 

[Raiden.Kaminari]

So basically you are giving new league players a big FU. Everyone has to start somewhere, if they were a previous league participant ok, I see the point but you will run off more than you keep by not allowing first year league players to advance with their teams.

Exactly the point of USTA management.

I can't find the e-mail, but I received one saying that in terms of self-rated players, the majority (I think the number was 99%) are honest. They are playing within their level.

Of the 1% that mis-rate, about 80% mis-rated only by .5. 18% mis-rated by 1.0. And only 2% were so out of level, it required suspensions for not following "fair play."

Personally, I like K-Swiss's Ultimate Tennis. When you're new to the program, you can not participate in the year end championships. You also get disqualified for under-rating very quickly

 

[kylebarendrick]

Self-rated players that can win matches at the Sectional/National level either self-rated too low or improved a lot during the season. If the former, then I agree they shouldn't be playing. If the latter, then I don't believe it is reasonable to stop them.

My proposal is to require self-rated players to play in more matches than computer rated players in order to become playoff eligible. I'd like to see at least 3 and would prefer something like 5. This helps to keep captains from hiding players during the season and provides enough results to generate a dynamic rating. I'd then create ESRs for self-rated players at the end of local league play and disqualify any over a certain threshold from the playoffs, without retroactively overturning previous matches.

 

[Cindysphinx]

What about 5.0s Cindy? We've actually got some computer rated 5.0s, and occasionally they can field two teams and have a league!

Not a single 5.0 woman in our league.

The 4.5 women were funny. One team was all 4.0s with a couple of 4.5s. The other team was all 4.5s with two 4.0s.

They played five times. Guess which team won?

Yeah. Being a 4.0 woman sounds like a total drag. I think I'll just stay in 3.0 where I have lots of company!

 

[Cindysphinx]

My proposal is to require self-rated players to play in more matches than computer rated players in order to become playoff eligible. I'd like to see at least 3 and would prefer something like 5. This helps to keep captains from hiding players during the season and provides enough results to generate a dynamic rating. I'd then create ESRs for self-rated players at the end of local league play and disqualify any over a certain threshold from the playoffs, without retroactively overturning previous matches.

Brilliant!

 

[Cindysphinx]

Cindy, I don't think you are skewed at all. In the Southern Section, the distribution is:

2.5 - 3%
3.0 - 33%
3.5 - 40%
4.0 - 19%
4.5+ - 8%

Your's seem to be in line.

In line with Southern, perhaps. In line with something that makes sense? I'm not so sure.

I mean, if you take the very worst 3.0 player in our league and line everyone up until you get to the top 3.5 player, you will have a line 713 people long. The range of ability across those two divisions (the divisions you will see competing against each other in 6.5 combo and 7.0 mixed) is way too big, IMHO. Heck, the range of ability between even the top 3.0 player and the bottom 3.0 player is hopelessly vast.

If the idea of having rating levels is to have people of similar ability playing each other, then you need to have more playing divisions where people tend to bunch up, or you need to tweak your formula to scoot some of these women up.

Especially when one of the biggest, fastest-growing leagues in a major metropolitan area can't even field a flight of 5.0 women and half of the women playing 4.5 are playing up.

 

[Raiden.Kaminari]

Self-rated players that can win matches at the Sectional/National level either self-rated too low or improved a lot during the season. If the former, then I agree they shouldn't be playing. If the latter, then I don't believe it is reasonable to stop them.

My proposal is to require self-rated players to play in more matches than computer rated players in order to become playoff eligible. I'd like to see at least 3 and would prefer something like 5. This helps to keep captains from hiding players during the season and provides enough results to generate a dynamic rating. I'd then create ESRs for self-rated players at the end of local league play and disqualify any over a certain threshold from the playoffs, without retroactively overturning previous matches.

For the adult season, in order to participate in post season, most LLAR (local league area rules) require you to play at least two matches. In order to participate in Nationals, you must have played at least 3 matches.

Some sections overturn matches when players get three strike DQ'd. At other sections, they keep the matches played, but tell the player they can't play at a particular NTRP level anymore.

What I've seen a few teams do is to self-rate one year, lose, and then suddenly go to Nationals the very next year. Talk about patience. People who cheat will always cheat.

 

[raiden031]

Good thoughts on preventing sandbaggers, especially self-rated ones. But nobody addressed my question of how they keep the ratings from skewing in one direction as a whole.

 

[raiden031]

I think the ratings in our area *are* quite skewed. Even the numbers show it. This spring, there were:

68 2.5s
400 3.0s
313 3.5s
137 4.0s
31 4.5s (with only 18 being actual 4.5s)

Shouldn't that distribution be a bit flatter? Shouldn't this be achieved by pushing some of the 3.0s and 3.5s up?

I also think it is weird that the ladies bunch up at 3.0 but the guys bunch up at 3.5 and 4.0.

Don't know if that is what you were driving at, Raiden, but there you have it.

Where did you get this information?

 

[cak]

Okay, I'm going to throw out the proposal I've thrown out before. To grow tennis, and especially to grow USTA tennis leagues, your first year of league play, until you get a computer rating, is free. That's right, your first year FREE!!! (Well, you have to join USTA, but otherwise, FREE!!!)

However, all players without computer ratings or ESRs cannot play in playoffs. They can play in league play, but not even local playoffs. And for most those players new to leagues, thats probably plenty. Yep, this would mean there probably will be somewhere around zero 2.5 teams eligible for playoffs, but I would say 2.5 teams shouldn't be in playoffs anyway. (I would even be amenible to no Nationals until 4.0, but then again, that's me.) If you happen to have a new player that played tons for your team (read, 5 matches or more) and really, really wants to play in playoffs, you pay the league fee at the end of the season, and voila, they pull up his or her ESR. Now this is where the sandbagging will be caught. If they played five 3.5 matches, and end up with an ESR that would put them in the 4.0 range, well then, they couldn't play in the playoffs for the 3.5 team. But for those 99% of the players who self rate fairly, this won't even come up.

What do you all think?

 

[kylebarendrick]

Your 5 matches and then use the ESR for playoffs approach is basically what I had in mind, so I think it sounds fine! I don't think league fees are much of a deterrent though. Heck, I pay more for court reservations and post-match snacks than I do for the league.

 

[kylebarendrick]

Good thoughts on preventing sandbaggers, especially self-rated ones. But nobody addressed my question of how they keep the ratings from skewing in one direction as a whole.

At this point I believe USTA is trying too hard not to move people up. The result is compressing players at the lower levels and leaving the higher levels depleted. The obvious first step is to eliminate the automatic ratings appeals. If you get bumped up then play a year at the next level!

Somehow I'd love for more people to get moved up/down each year, so we can get away from the idea of being stuck at a level. I'd rather people see getting bumped from 3.5 to 4.0 as a promotion (like going from A to AA baseball) rather than a death sentence. Hopefully if it is more common then people will worry about it less.

There will always be people who want to cheat for the sake of winning. I'm convinced that can't be stopped and the system shouldn't be designed for them.

 

[Cindysphinx]

Where did you get this information?

In Tennislink. For each level, you can click "Player count."

I should add that my numbers are the numbers of people playing at a particular level. So a 2.5 who plays 2.5 and 3.0 would be double-counted.

It would be a lot more work to subtract out people who are playing up, so I didn't do it.

 

[tfm1973]

Cindy - It's slow here at the office and you piqued my curiosity. I did some more number crunching for you. In the Mid Atlantic Section and Montgomery County Distric for the Adult Leagues:

1605 total players (roughly because some playing more than 1 level) -
656 Men and 949 Women

2.5: 14 men (2%) 68 women (7%)
3.0: 109 men (16%) 400 women (42%)
3.5: 205 men (31%) 313 women (33%)
4.0: 233 men (35%) 137 women (14%)
4.5: 67 men (10%) 31 women (3%)
5.0: 28 men (4%)

For comparison here's Northern Virginia:

2171 total players (no 2.5 players or 5.0 players in their Adult Indoor League)
921 Men and 1250 Women

3.0: 154 men (16%) 455 women (36%)
3.5: 325 men (35%) 391 women (31%)
4.0: 301 men (32%) 278 women (22%)
4.5: 141 men (15%) 126 women (10%)

 

[Topaz]

Cindy - It's slow here at the office and you piqued my curiosity. I did some more number crunching for you. In the Mid Atlantic Section and Montgomery County Distric for the Adult Leagues:

1605 total players (roughly because some playing more than 1 level) -
656 Men and 949 Women

2.5: 14 men (2%) 68 women (7%)
3.0: 109 men (16%) 400 women (42%)
3.5: 205 men (31%) 313 women (33%)
4.0: 233 men (35%) 137 women (14%)
4.5: 67 men (10%) 31 women (3%)
5.0: 28 men (4%)

For comparison here's Northern Virginia:

2171 total players (no 2.5 players or 5.0 players in their Adult Indoor League)
921 Men and 1250 Women

3.0: 154 men (16%) 455 women (36%)
3.5: 325 men (35%) 391 women (31%)
4.0: 301 men (32%) 278 women (22%)
4.5: 141 men (15%) 126 women (10%)

TFM, there are actually 2.5 and 5.0 players in the NOVA indoor league, there just aren't enough to form their own leagues. The 2.5s played up at 3.0, and I'm not sure what the 5.0s did (though, are you sure there weren't some 5.0 teams? I seem to remember seeing them.)

 

[tfm1973]

the 5.0's in NOVA? they changed their names, self rated and are kicking my balls around in 4.0 leagues.

 

[JavierLW]

I think the ratings in our area *are* quite skewed. Even the numbers show it. This spring, there were:

68 2.5s
400 3.0s
313 3.5s
137 4.0s
31 4.5s (with only 18 being actual 4.5s)

Shouldn't that distribution be a bit flatter? Shouldn't this be achieved by pushing some of the 3.0s and 3.5s up?

I also think it is weird that the ladies bunch up at 3.0 but the guys bunch up at 3.5 and 4.0.

Don't know if that is what you were driving at, Raiden, but there you have it.

It shouldnt necessarily be any flatter because among adults, there defaintely are more 3.0, 3.5, and 4.0 players than any other player.

This is because if you have some athletic ability you automatically many times can be a good 3.0 or 3.5, and can possibly advance to 4.0

IT'S MUCH HARDER TO ADVANCE TO 4.5 FROM 4.0 (or even higher).

It's much harder to get that good, so there are less of those players. And if your area is smaller, you may not have any. (like in your case, you dont seem to have any 5.0, 5.5, or Open players, my area has like a handful)

I dont think artificially pushing people up really is necessary. The only reason to do that would to be to make it so more teams could advance and that's not really a reason to do that.

Anyways those numbers are normal. It's a "skill" level after all, and there is no way that it should necessarily be flat.

(think of it like this, on the very top, 5.5 is close to "Open" level. Are you going to really insist that there should be as many players in "Open" level as you have in 3.0???)

 

[JRstriker12]

Just a thought,

How about requiring new players to be rated by a USTA pro when they enter the league?

Do something like set up a new player orientation before the season opens and provide free rating services. I think it would be kind of hard for a 4.5 or 4.0 player fake being a 3.0 or 3.5 player without the pro noticing. Also, I would guess some of the higher rated players may be know to some of the pros in the area, so they may catch a sandbagger or two. If they miss the session then that player can pay for thier own rating session with a pro.

I know one pro in Reston that does free ratings as a way of getting new business. That's one angle of getting the pros to do the ratings.

 

[JavierLW]

Just a thought,

How about requiring new players to be rated by a USTA pro when they enter the league?

Do something like set up a new player orientation before the season opens and provide free rating services. I think it would be kind of hard for a 4.5 or 4.0 player fake being a 3.0 or 3.5 player without the pro noticing. Also, I would guess some of the higher rated players may be know to some of the pros in the area, so they may catch a sandbagger or two. If they miss the session then that player can pay for thier own rating session with a pro.

I know one pro in Reston that does free ratings as a way of getting new business. That's one angle of getting the pros to do the ratings.

They used to do that, it's was called a "rating clinic".

It was easy to "fake" your rating if you really wanted to do so. There are too many players in relation to the amount of pros on there, so many times you end up having 4 random players on a court in a short 30 minute session, just to make sure everyone gets rated.

And the USTA wants players to play so I think that's why they did away with the self-rating requirement. It was just a needless extra chore that you had to perform that really didnt mean a whole lot anyway.

The only thing that I think that is missed from not having these was it was a great chance for interested players to get together with captains who were looking for players, as well as get some orientation into the league. (now the orientation part is totally up to the captain)

 

[smiley74]

I posted before about my idea for this but will repeat it again......

What about having like a gold, silver, bronze, and novice level?

Novice level would be for beginners(2.0 -2.5). After you win so many matches, you are automatically bumped up to the bronze level (3.0-3.5). After you win blank number of matches, you are moved up again to the silver(4.0). Same thing there and you are moved up to the gold (4.5 +).

This would keep the playing field level because you wouldn't be allowed to stay in one division and just keep winning. Plus, it would give true beginners a shot. A lot of people don't compete because they know that everyone plays up a level. So, if you are like me and a true beginner, you get bumped to 3.0 or play 2.5s who are really 3.0s.

Plus, you could add that only the gold level gets to play at Nationals. Thus, people have incentive to get to that level and only the cream of the crop get to attend. Golds would have to qualify by winning a certain number of matches and or doing well at Sectionals.

The silvers and bronzes would get to play at Sectionals in their own division along with the golds. So, silvers and bronzes could be sectional champions at their level.

I don't know. Just a thought.....

 

[cak]

Smiley74, I like the ideas that winners move up. There was a local rec league I heard about in San Diego where there were something like four seasons a year. And if you won 66% of your matches you moved up for the next season, but if you lost 66% of your matches you moved down. The ladies who played in this league loved the competitive matches. You would only have three months of winning everything, and only three months of losing everything. Usually you'd end up in a league where you won about half of your matches.

 

[CAM178]

I do not like the self-rating system that the USTA has adopted. That implies honesty, and some people wil always tell a little white lie when it comes to their rating. Ratings clinics used to be mandatory in order to get into a league. That was more representative. That being said, I cheated the system once upon a time. Got rated as a 3.5, and I'm not proud of it. I did it as a favor for a buddy's team that wanted me. I barely got the rating, as I did something during the clinic that caused all of the raters to ask me to 'stay after class'.

I like the K-Swiss league level system, as it is based on wins. If you win, you move up. You lose a lot, you go down. Same as with Atlanta's ALTA.

 

[Raiden.Kaminari]

Okay, I'm going to throw out the proposal I've thrown out before. To grow tennis, and especially to grow USTA tennis leagues, your first year of league play, until you get a computer rating, is free. That's right, your first year FREE!!! (Well, you have to join USTA, but otherwise, FREE!!!)

However, all players without computer ratings or ESRs cannot play in playoffs. They can play in league play, but not even local playoffs. And for most those players new to leagues, thats probably plenty. Yep, this would mean there probably will be somewhere around zero 2.5 teams eligible for playoffs, but I would say 2.5 teams shouldn't be in playoffs anyway. (I would even be amenible to no Nationals until 4.0, but then again, that's me.) If you happen to have a new player that played tons for your team (read, 5 matches or more) and really, really wants to play in playoffs, you pay the league fee at the end of the season, and voila, they pull up his or her ESR. Now this is where the sandbagging will be caught. If they played five 3.5 matches, and end up with an ESR that would put them in the 4.0 range, well then, they couldn't play in the playoffs for the 3.5 team. But for those 99% of the players who self rate fairly, this won't even come up.

What do you all think?

It's a fine idea and I think several others have brought it up before.

The only problem is that an ESR isn't really valid because while it may apply to your local area, it may not apply to the section (until Districts are played). And the year end rating won't be known until the results of Nationals are in.

Depending on how your Section did, certain levels will get a boost in their year end rating (especially if the Section made the final four), while other levels will remain the same or drop (depending on how poorly the team did at Nationals).