Share your WTN - Crowd sourcing the NTRP to WTN mapping

Roforot

Hall of Fame
At a recent tournament, I asked about how they set up seeds b/c I was surprised we were seeded ahead of a team who beat us.
Apparently, the website assigns seeds based on WTN #s... well seeds are not that big of a deal for amateurs, but I was surprised that
the WTN number actually has even that level of relevance.
 

ChaelAZ

G.O.A.T.
Since this was back in July, was curious to see what kind of decay, drift, or outright changes were made with anything, but really not much. I haven't played a ranked match since the end of March (and think I am pretty much done with USTA leagues overall).

WTNs 31.4 (same)
WTNd 30.2 (same)
UTRs 4.82 (bit of change)
UTRd 5.39 (bit of change)

Still have it in my head to play some UTR tourneys or such. Will see.


No play since March 2023 and let my USTA membership lapse, but everything is still about the same:

USTA 4.0
TR - 3.55
WTNs 31.43
WTNd 31.19
UTRs - 5.06
UTRd - 5.52

Only playing socially at the moment with our 4.0 team players. Kinda wish I had played another season. Most the guys are going to get bumped to 4.5.
 

InsuranceMan

Hall of Fame
Not really. It doesn’t update enough to be remotely accurate.
You might be right, my mind immediately when to a demographic of junior/college/adults who play UTR but not everybody is that active with as many results on UTR. NTRP is garbage imo and judging off what I’ve seen this thread and my own life WTN is also not good. I’m not saying I like UTR, but with the caveat that there’s plenty of active, relevant rated match play I feel like it’s the best
 

time_fly

Hall of Fame
No play since March 2023 and let my USTA membership lapse, but everything is still about the same:

USTA 4.0
TR - 3.55
WTNs 31.43
WTNd 31.19
UTRs - 5.06
UTRd - 5.52

Only playing socially at the moment with our 4.0 team players. Kinda wish I had played another season. Most the guys are going to get bumped to 4.5.

Interesting. Me currently:
USTA 3.5C
TR - 3.55
WTNs 30.63
WTNd 30.17
UTRs 4.52
UTRd 4.89

Seems like UTR vs. WTN calibration may depend on region. Or it may just be that I play a lot of mixed doubles, which tends to kill UTR.
 

CiscoPC600

Hall of Fame
I am not sure if this has been mentioned in prior posts, but I spoke to a tournament director for one of the largest tournaments in San Diego, and she said that WTN ratings with high confidence are the only ratings used for seedings. Apparently even medium confidence are not used because they're not reliable.

I'm curious if that changes this thread's perspective on the ratings reliability. Not sure how many people have been mentioning their respective confidence levels.
 

Moon Shooter

Hall of Fame
I am not sure if this has been mentioned in prior posts, but I spoke to a tournament director for one of the largest tournaments in San Diego, and she said that WTN ratings with high confidence are the only ratings used for seedings. Apparently even medium confidence are not used because they're not reliable.

I'm curious if that changes this thread's perspective on the ratings reliability. Not sure how many people have been mentioning their respective confidence levels.
Do you mean the blue check? If so then not that doesn’t change anything. WTN was bad at the start but then USTA seemed all at once to do a massive rating change for people that made it go from bad to completely worthless a while back. It seems they intentionally made it inaccurate.

I am not even sure it counts all the games as opposed to just the sets.
 

schmke

Legend
I am not even sure it counts all the games as opposed to just the sets.
I attended a webinar that covered WTN and the explicitly said they only use sets, not games. Their rationale is that in testing it was as or more accurate that way. No details provided on the result of the testing or what the alternative was so we'll just have to take their word, or rather the ITF's word.
 

CiscoPC600

Hall of Fame
Do you mean the blue check? If so then not that doesn’t change anything. WTN was bad at the start but then USTA seemed all at once to do a massive rating change for people that made it go from bad to completely worthless a while back. It seems they intentionally made it inaccurate.

I am not even sure it counts all the games as opposed to just the sets.

Yes, blue check and it says "high confidence" right next to it. If you have a red check then it says "medium confidence" and a gray check is "low confidence."

And at least in So Cal tournaments, high confidence ratings is the sole factor in determining tournament seeds.
 

TennisOTM

Professional
I attended a webinar that covered WTN and the explicitly said they only use sets, not games. Their rationale is that in testing it was as or more accurate that way. No details provided on the result of the testing or what the alternative was so we'll just have to take their word, or rather the ITF's word.
Wow, hard to believe that ignoring the games data would make it *more* accurate. I can believe that, for a group of players who have all played lots of matches, the sets-only data would converge pretty close to a result that uses games data. But when the network includes thousands of players who have only played a few matches, like adult rec players, I would think that games scores would be critical for improving the estimates.

Also, do they treat match tie-breaks as a "set" that's worth the same as a full set? I wonder if that choice was part of their testing...
 

ohplease

Professional
Wow, hard to believe that ignoring the games data would make it *more* accurate. I can believe that, for a group of players who have all played lots of matches, the sets-only data would converge pretty close to a result that uses games data. But when the network includes thousands of players who have only played a few matches, like adult rec players, I would think that games scores would be critical for improving the estimates.

Also, do they treat match tie-breaks as a "set" that's worth the same as a full set? I wonder if that choice was part of their testing...

Game counts are unfortunately easily manipulated. Sets are better but again people can drop sets and still easily win the super tie. Even match results are suspect because teams will dump entire leagues just to keep their ratings (for example - they really want to go to nationals in the 55s, so they play 18s and lose badly every chance they get). Or the folks who dump everywhere so they can play on mixed superteams.

It wouldn't be that hard to attach a probability of tomfoolery to scores and just throw out the ones that are clearly not real. Like, if people are going to spend money to go to districts, sectionals, nationals, etc. - those are clearly more likely to be real and you can derive everyone else's ratings from the network of matches from when those teams actually were trying.

People are already doing stuff like this and probably much more to detect betting and match fixing on the tours - but hoo yeah rec league shikata ga nai
 

TennisOTM

Professional
Game counts are unfortunately easily manipulated. Sets are better but again people can drop sets and still easily win the super tie. Even match results are suspect because teams will dump entire leagues just to keep their ratings (for example - they really want to go to nationals in the 55s, so they play 18s and lose badly every chance they get). Or the folks who dump everywhere so they can play on mixed superteams.

It wouldn't be that hard to attach a probability of tomfoolery to scores and just throw out the ones that are clearly not real. Like, if people are going to spend money to go to districts, sectionals, nationals, etc. - those are clearly more likely to be real and you can derive everyone else's ratings from the network of matches from when those teams actually were trying.

People are already doing stuff like this and probably much more to detect betting and match fixing on the tours - but hoo yeah rec league shikata ga nai
Players trying to manipulate things by playing poorly on purpose are going to be a problem for any rating system. Ignoring the game score doesn't really solve it. If it works to give more weight to winning sets and matches then that could be done without ignoring the game score completely. The game score, set score, and match winner could all be factored in to the calculations. It's hard to believe the games component doesn't help at all, especially for newer players with only a few match results. The vast majority of players are trying to win every game.
 

Moon Shooter

Hall of Fame
I attended a webinar that covered WTN and the explicitly said they only use sets, not games. Their rationale is that in testing it was as or more accurate that way. No details provided on the result of the testing or what the alternative was so we'll just have to take their word, or rather the ITF's word.

No way they used typical adult rec tennis players. It is interesting how their secret testing was different then USTA’s and UTR’s and even testing here where even TR is more accurate then WTN. It’s unbelievable. Why can’t anyone in tennis just implement a decent rating system like every online chess website does?
 

travlerajm

Talk Tennis Guru
I wonder if USTA will eventually use WTN for bumps/demotions.
WTN has picked up my mixed match result from this weekend and adjusted my rating accordingly.

Meanwhile, no sign of the match on TR or UTR.

Seems like USTA is trying to monopolize the use of the data to get more traction with WTN.

If they continue to freeze out TR and UTR, I imagine that WTN will gain ground in the rating credibility and trust race.
 

CiscoPC600

Hall of Fame
WTN has picked up my mixed match result from this weekend and adjusted my rating accordingly.

Meanwhile, no sign of the match on TR or UTR.

Seems like USTA is trying to monopolize the use of the data to get more traction with WTN.

If they continue to freeze out TR and UTR, I imagine that WTN will gain ground in the rating credibility and trust race.
Yeah, and so far it seems mostly reliable for me to gauge someone if they have a "high confidence" rating. There's still a few outliers I catch (3.5C players with better WTN than some 4.0/4.5), but I imagine over the next year, those will disappear as the data size grows for each and every player.
 

schmke

Legend
WTN has picked up my mixed match result from this weekend and adjusted my rating accordingly.

Meanwhile, no sign of the match on TR or UTR.

Seems like USTA is trying to monopolize the use of the data to get more traction with WTN.

If they continue to freeze out TR and UTR, I imagine that WTN will gain ground in the rating credibility and trust race.
I don't know that it will gain credibility, but it may become what is used simply because the other sites aren't updating. Once people look at it more and see some of the oddities it spits out, it could lose credibility.
 

schmke

Legend
Yeah, and so far it seems mostly reliable for me to gauge someone if they have a "high confidence" rating. There's still a few outliers I catch (3.5C players with better WTN than some 4.0/4.5), but I imagine over the next year, those will disappear as the data size grows for each and every player.
When they started they had 5 years of data used from what they said. And two years in that is now 7 years. Yes, more data is better, but lack of data should not be an excuse for what they are publishing now.
 

travlerajm

Talk Tennis Guru
I don't know that it will gain credibility, but it may become what is used simply because the other sites aren't updating. Once people look at it more and see some of the oddities it spits out, it could lose credibility.
The WTN ratings seemed to make more sense before they did the “recal” or whatever they actually did.

Seems like they tried to force fit adjustments based on age and gender instead of just keeping the algo simple and letting the results sort everything out.

In effect, it seems WTN is trying to outsmart itself with forced constraints, much like UTR is with its self-defeating arbitrary exclusions based on match score and rating gap.
 

CiscoPC600

Hall of Fame
When they started they had 5 years of data used from what they said. And two years in that is now 7 years. Yes, more data is better, but lack of data should not be an excuse for what they are publishing now.
That's fair. But not everyone has been playing the past 7 years.
 

TennisOTM

Professional
Can someone explain *how* USTA would be able to "freeze out" sites like TR and UTR from using their data? Are they taking some kind of legal cease-and-desist action?

If not, and they just changed their data to make it harder to scrape, then that seems like a solvable problem. As long as the match result data are publicly out there, there must be a way to collect it - computer / data guys can figure anything out.
 

Moon Shooter

Hall of Fame
That's fair. But not everyone has been playing the past 7 years.
Not everyone will be playing 7 years from now. There will always be people going in and out of the adult rec ranks.

As I recall (and I may be misremembering this) the median usta adult rec player plays about 6 matches a year. When you only count sets, start off with crazy self imposed ratings (not based on actual results) those rating will never correct to anything useful. Older 3.0 male self players seem to be assigned a number like 34. Younger 3.0 male self rates seem to be assigned a number like 28. But it may be different depending on when they started.
 

Roforot

Hall of Fame
The nice thing about wtn and utr is that they separate singles and doubles. I know it would make data a bitmore tricky but having a separate ranking makes sense. I could play 4.0 doubles fine but 3.5 singles is tough especially i. Tournaments. Granted utr is gone loco raising my singles to 6 and dropping my doubles to 5.3… after we won a doubles tournament…
 

silverwyvern4

Semi-Pro
When they started they had 5 years of data used from what they said. And two years in that is now 7 years. Yes, more data is better, but lack of data should not be an excuse for what they are publishing now.
It's like an unknown politician. Right now it was low favorable and unfavorable rating and most people have no opinion. Once they get to know WTN, it's unfavorable polling will rise
 

silverwyvern4

Semi-Pro
Can someone explain *how* USTA would be able to "freeze out" sites like TR and UTR from using their data? Are they taking some kind of legal cease-and-desist action?

If not, and they just changed their data to make it harder to scrape, then that seems like a solvable problem. As long as the match result data are publicly out there, there must be a way to collect it - computer / data guys can figure anything out.
Tennis record is still fetching new match results every week
 

CiscoPC600

Hall of Fame
Yeah, I'm not convinced that USTA is actually doing anything right now to hinder the non-WTN rating sites. But for those who think they are, what exactly would/could they be doing?
They can send cease and desist letters and initiate lawsuits. Doubt the latter but the former could scare off TR or UTR.

From a tech perspective, they could have finally found a way to make data scraping harder or more expensive.
 

Roforot

Hall of Fame
They can send cease and desist letters and initiate lawsuits. Doubt the latter but the former could scare off TR or UTR.

From a tech perspective, they could have finally found a way to make data scraping harder or more expensive.
If the usta really wanted to shut down, they could release the dynamic ratings theyve been using. The fear I suppose is that once ppl get their ratings to be at a certain point, they may stop playing so they end the year at that rank. I feel they could give it out at certain intervals To avoid this
 

Moon Shooter

Hall of Fame
Yeah, and so far it seems mostly reliable for me to gauge someone if they have a "high confidence" rating. There's still a few outliers I catch (3.5C players with better WTN than some 4.0/4.5), but I imagine over the next year, those will disappear as the data size grows for each and every player.

What WTN ranges do 3.5 women have in your area? In my area the 3.5 women have 27-32 WTNS which is about the same as upper 3.5 to 4.0 men.

The ratings system should be judged as accurate, or not, based on their ability to predict outcomes. WTN came in dead last at 61% accuracy in this test:
Which is the only objective test I have seen trying to measure accuracy between these rating systems.

61% might seem pretty good but consider that flipping a coin should get you about 50%.
 

CiscoPC600

Hall of Fame
What WTN ranges do 3.5 women have in your area? In my area the 3.5 women have 27-32 WTNS which is about the same as upper 3.5 to 4.0 men.

The ratings system should be judged as accurate, or not, based on their ability to predict outcomes. WTN came in dead last at 61% accuracy in this test:
Which is the only objective test I have seen trying to measure accuracy between these rating systems.

61% might seem pretty good but consider that flipping a coin should get you about 50%.
I took a quick look at the 4 women in the semis of large tournament here and it was: 32, 32, 31, and 29.
 

SlangB

New User
I received an email from the USTA about creating an ITF account earlier this week, so I signed up and played around with it a bit. One aspect I like is the ability to create "watch lists", which allows you to create fifty different groups of up to 200 players per group. It's similar to UTR's groups except it has built in min, max, and average ratings for singles; I don't see any way to toggle it for doubles. There's also an option to export to CSV, which pretty easily solves any problems for figuring out teamwide doubles ratings. I can't seem to add myself to these watchlists, but I'm hoping that's unintended and will be fixed

 
Top