The ITF has changed the WTN algorithm, is it any better?

[TennisBro]

Seems that serious topics with facts in are uncomfortable for the boards here

Here's another fact-my 15 years old junior ITF son, whose WTN's at 20.2 and UTR at 10.5, has just beaten a 1,100th ranked ATP player with WTN at 6.5 in a M15 tournament. Match 7:5, 6:1 and the Argentinean opponent broke his racket and got a code violation with a fine alongside; so, not a fluke win for my teenage son and not the first one either.

The truth is that junior players training and playing in certain parts of the world are discriminated. They have to make giant leaps forward to have their WTN/UTR corrected. The system's been set to favor the areas with strong market for tennis.

 

[schmke]

Seems that serious topics with facts in are uncomfortable for the boards here

Here's another fact-my 15 years old junior ITF son, whose WTN's at 20.2 and UTR at 10.5, has just beaten a 1,100th ranked ATP player with WTN at 6.5 in a M15 tournament. Match 7:5, 6:1 and the Argentinean opponent broke his racket and got a code violation with a fine alongside; so, not a fluke win for my teenage son and not the first one either.

The truth is that junior players training and playing in certain parts of the world are discriminated. They have to make giant leaps forward to have their WTN/UTR corrected. The system's been set to favor the areas with strong market for tennis.

I wouldn't go so far as to say players are discriminated against, but my observation is that WTN has a cohort (as they call it) problem.

The nature of how tennis is played results in many cohorts, or islands, of players that play each other but don't play other cohorts, and with no match play between them there is no algorithmic way to ensure relative ratings between the cohorts are accurate. The cohorts can be regional, or organizational.

For example, WTN calculates ratings for USTA League and Tournament players as well as juniors and pros. For the most part, these groups of players don't play each other, so the ratings within a group may be relatively accurate (debatable, but work with me), but the ratings between the groups could be off. In the specific case of USTA players, the USTA/ITF did a big adjustment last year effectively moving a USTA cohort to higher numbered ratings as they didn't think the relative ratings with juniors were accurate, e.g. a USTA League WTN 20 was not as good a player as a junior WTN 20, so the USTA cohort was largely all moved higher about 5 points.

This can obviously happen regionally too even within juniors that play internationally as the connection between the regional cohorts may be small, and if the algorithm isn't set up to give the matches between cohorts extra weight in doing cohort adjustments, it can take a long time for there to be sufficient play to get the relative ratings accurate.

 

[TennisBro]

@schmke , overall view on the topic shows your deep understanding of WTN and UTR in tennis; your reaction to my post offers a degree of clarity.

I wouldn't go so far as to say players are discriminated against, but my observation is that WTN has a cohort (as they call it) problem.

The nature of how tennis is played results in many cohorts, or islands, of players that play each other but don't play other cohorts, and with no match play between them there is no algorithmic way to ensure relative ratings between the cohorts are accurate. The cohorts can be regional, or organizational.

Venting frustration sure isn't popular, using strong wordings may be extreme for some. However, if "cohorts" differentiate right of the start with the sport and junior athletes and place them into "islands" while putting higher value to their favorite piece of land and people, they are nothing but racial bigots. Haven't we had that kind of attitude in the international footbal (soccer) a few decades back?

For example, WTN calculates ratings for USTA League and Tournament players as well as juniors and pros. For the most part, these groups of players don't play each other, so the ratings within a group may be relatively accurate (debatable, but work with me), but the ratings between the groups could be off. In the specific case of USTA players, the USTA/ITF did a big adjustment last year effectively moving a USTA cohort to higher numbered ratings as they didn't think the relative ratings with juniors were accurate, e.g. a USTA League WTN 20 was not as good a player as a junior WTN 20, so the USTA cohort was largely all moved higher about 5 points.

The US colleges and recruiters/sponsors from NY to LA are heavy on unis applicants' UTR which assists them in their decision making process who'd be presented with D1/D2 and full/partial sponsorship opportunities. Those offers come out of juniors' effort on their "islands" and from those "cohorts" too. While the stats of predominantly foreign D1/D2 tennis players in the US colleges are commendable, a data of which nations those athletic students come from is to be looked at as is from which "islands" and perhaps backgrounds/families the young and lucky ones are from.

This can obviously happen regionally too even within juniors that play internationally as the connection between the regional cohorts may be small, and if the algorithm isn't set up to give the matches between cohorts extra weight in doing cohort adjustments, it can take a long time for there to be sufficient play to get the relative ratings accurate.

Very much so. This is evident in NA, within the US/Canada and in between the US and Canada as well as it is within Asia and its nations too. That "long time" for adjustments in the more precise grading of those fine players' efforts causes the 17-18 year olds lose their opportunities and forces some of them, when younger, to move to more correct "islands" and more powerful "cohorts" in order to succeed. That's how talent/skill is lost when youngsters are forced to make their choices. Such practices increase the monopoly of the young talented tennis players to develop in areas of the powers' preferences rather than in places where tennis is loved unconditionally.

The world of tennis ought to be a lot more open to all regardless their "islands" or "cohorts".

 

[Moon Shooter]

The Science Behind World Tennis Number

Discover everything you need to know about World Tennis Number to grow your game and grow our sport. Find opponents with Game zONe and access statistics.

worldtennisnumber.com

"Match results are analysed at set level, meaning the algorithm takes into account each individual set as its own result. Example If a match ends two sets to one, then the system will update the winning player’s Number with two set ‘wins’ and one set ‘loss’. Even if a player doesn’t win the match any sets they have won will count towards their Number."

https://www.lta.org.uk/4ad8ae/siteassets/about-lta/file/lta-itf-wtn-principles-guide-a5.pdf

Has anyone confirmed WTN actually considers games and not just set scores? I have never heard anyone from ITF or WTN definitively make it clear that a 7-6 7-6 score would be counted differently then a 6-0 6-0 score. That they are leaving so much relevant data out of the algorithm may help explain why this number seems so useless.

 

[schmke]

"Match results are analysed at set level, meaning the algorithm takes into account each individual set as its own result. Example If a match ends two sets to one, then the system will update the winning player’s Number with two set ‘wins’ and one set ‘loss’. Even if a player doesn’t win the match any sets they have won will count towards their Number."

https://www.lta.org.uk/4ad8ae/siteassets/about-lta/file/lta-itf-wtn-principles-guide-a5.pdf

Has anyone confirmed WTN actually considers games and not just set scores? I have never heard anyone from ITF or WTN definitively make it clear that a 7-6 7-6 score would be counted differently then a 6-0 6-0 score. That they are leaving so much relevant data out of the algorithm may help explain why this number seems so useless.

My observation with WTN is if you win, your rating goes up, if you lose, it goes down. It doesn't seem to give credit for losing close to stronger players.

The amount you move does seem to be somewhat dependent on sets won/lost and the strength of the opponent I think matters too, but those factors seem smaller than you might think.

For example, I had a doubles loss, 6-4,7-5, where the opponents were 2.65 better than my partner and I, you'd think a good result, but my WTN got 0.2 worse as a result. The another 6-4,7-6 loss to opponents 2.3 better, and my WTN got another 0.2 worse. So competitive matches against better opponents still moved my WTN in a negative direction, more than I at least would think. It seems to simply say I lost 2 sets so my WTN gets worse.

 

[Moon Shooter]

My observation with WTN is if you win, your rating goes up, if you lose, it goes down. It doesn't seem to give credit for losing close to stronger players.

The amount you move does seem to be somewhat dependent on sets won/lost and the strength of the opponent I think matters too, but those factors seem smaller than you might think.

For example, I had a doubles loss, 6-4,7-5, where the opponents were 2.65 better than my partner and I, you'd think a good result, but my WTN got 0.2 worse as a result. The another 6-4,7-6 loss to opponents 2.3 better, and my WTN got another 0.2 worse. So competitive matches against better opponents still moved my WTN in a negative direction, more than I at least would think. It seems to simply say I lost 2 sets so my WTN gets worse.

Well that would help explain some of WTN's woes. But what explains why they would *choose* to only use sets? I mean it is obvious to just about anyone that only looking at sets ignores too much useful data. (and likely over-weights match "tiebreakers")

Is it the same guy that decided the higher rated player would have a lower number? Why can't tennis get a rating system that does not seem to make *intentionally* bad decisions?

 

[schmke]

Well that would help explain some of WTN's woes. But what explains why they would *choose* to only use sets? I mean it is obvious to just about anyone that only looking at sets ignores too much useful data. (and likely over-weights match "tiebreakers")

Is it the same guy that decided the higher rated player would have a lower number? Why can't tennis get a rating system that does not seem to make *intentionally* bad decisions?

Why did they do "reverse" order? I don't know for sure, but I'm guessing:

3 - This is ITF run (along with other governing bodies) and the original International Tennis Number (ITN) was 10 to 1, 1 being the best, so this is consistent with it
2 - The ITF thinks ratings should mirror rankings, e.g. you strive to be #1, and that makes it easier to understand
1 - Having the "best" be a fixed value, 1.0, gives a clear reference point. A rating system going the other way is in theory unbounded on the top-end

Are any of those good reasons to do it? Probably not. #3 is just sticking to something to be consistent with something no one really used. I think rankings and ratings often get confused and it may actually be better to not try to have them be similar to show they are distinct. And the problem with a fixed reference point of 1.0 is that that implies everyone else has to scale up or down as the best player is better or worse than 1.0, it handcuffs the algorithm.

 

[Moon Shooter]

Why did they do "reverse" order? I don't know for sure, but I'm guessing:

3 - This is ITF run (along with other governing bodies) and the original International Tennis Number (ITN) was 10 to 1, 1 being the best, so this is consistent with it
2 - The ITF thinks ratings should mirror rankings, e.g. you strive to be #1, and that makes it easier to understand
1 - Having the "best" be a fixed value, 1.0, gives a clear reference point. A rating system going the other way is in theory unbounded on the top-end

Are any of those good reasons to do it? Probably not. #3 is just sticking to something to be consistent with something no one really used. I think rankings and ratings often get confused and it may actually be better to not try to have them be similar to show they are distinct. And the problem with a fixed reference point of 1.0 is that that implies everyone else has to scale up or down as the best player is better or worse than 1.0, it handcuffs the algorithm.

LOL thanks for numbering the points you make. As for your third point (which, of course, is labeled 1) does the best player have a rating of 1.00? It seems they just sort of punt on the whole pro ratings.

When UTR WTN or any rating system has a ceiling and a floor it makes it less helpful. One of best thing about chess ratings (which have no ceiling) was you knew that if you played someone say 200 points higher then you then they should win about 75% of the games. If you did better then that then those matches would increase your rating if you did worse then you would lose ratings points. But if you have a ceiling and floor you cant do that. You always have to reshuffle everyone inside of the arbitrary cap and floor and it is always unclear what 3 points of rating difference is even supposed to signify.

 

[TennisBro]

My initial look is that it has changed, but I'm not sure it improved. More details on my blog, but personally and for those players I've partnered and played against, it simply appears all the rating got higher (worse rating) but sort of uniformly so. For example my WTNd went from around 22 to around 29 and my partners/opponents went up similarly. A teammate had been in the mid/high teens and is now at 25. And relative ratings between men and women are not noticeably different, meaning it is still broken.

Given the recent results, let me take another look at this issue from a prospective of a 15-year-old junior ITF player who competes in ATP Futures as well as junior ITF around the world and who plays with a large variety of players.

So, my teenage son beat a couple of players in M15 two weeks ago. One of those players in the first round of qualies was 1,100 ATP ranked one at way over UTR 13 and with WTN of 6.5. The boy's WTN jumped from 20.32 (two days ago) to 19.3 alright but his UTR hasn't significantly changed at all. I've just been told by experienced national pro coaching staff that those UTR ratings of 10.6 don't change much when playing (and winning with) so much higher UTR rated players that are larger than 12 UTR. Anotherwords, the 10.6 UTR teenager would have to play, and win, against UTR 11 players in order to faster improve his UTR rating. So, the logic in the UTR system appears to be that UTR 10.6 cannot quicker improve beating UTR 13-UTR 14 rated players but winning against UTR 11-UTR 12 players. How does that sound, for a professional rating system, to you guys?

Anyhow, the system is clearly "broken" and split into those suggested "islands" not only by location but rating categories players ought to consider matching up against if they want to improve their UTRs.

I loathe the people in the office that administer the kind of system which does not value tennis players' effort accordingly and which put some juniors in situations that take opportunities such as sponsorships or scholarships away from them.

 

[Moon Shooter]

Given the recent results, let me take another look at this issue from a prospective of a 15-year-old junior ITF player who competes in ATP Futures as well as junior ITF around the world and who plays with a large variety of players.

So, my teenage son beat a couple of players in M15 two weeks ago. One of those players in the first round of qualies was 1,100 ATP ranked one at way over UTR 13 and with WTN of 6.5. The boy's WTN jumped from 20.32 (two days ago) to 19.3 alright but his UTR hasn't significantly changed at all. I've just been told by experienced national pro coaching staff that those UTR ratings of 10.6 don't change much when playing (and winning with) so much higher UTR rated players that are larger than 12 UTR. Anotherwords, the 10.6 UTR teenager would have to play, and win, against UTR 11 players in order to faster improve his UTR rating. So, the logic in the UTR system appears to be that UTR 10.6 cannot quicker improve beating UTR 13-UTR 14 rated players but winning against UTR 11-UTR 12 players. How does that sound, for a professional rating system, to you guys?

No this does not sound right. If the player who is over 2.00 rating points higher wins the match then it wiill have no effect. But if the lower rated player actually wins the match then it will be counted. How much his rating will change depends on many factors - many of which have nothing to do with that particular match.

WTN is a complete mystery. It is unclear whether WTN considers all the games in a match or just who won each set. So a 6-0 6-0 might count the same as a 7-5, 7-6.

Anyhow, the system is clearly "broken" and split into those suggested "islands" not only by location but rating categories players ought to consider matching up against if they want to improve their UTRs.

I loathe the people in the office that administer the kind of system which does not value tennis players' effort accordingly and which put some juniors in situations that take opportunities such as sponsorships or scholarships away from them.

If you think some areas have over rated players then have your son play there. He will gain rating points until their rating accurately reflects their skill level.

Just like if they had true coed tournaments I think the men playing against women in my area would improve their wtn rating. It would be a bonanza for rating improvement for a time but then it would equal out.

 

[TennisBro]

No this does not sound right. If the player who is over 2.00 rating points higher wins the match then it wiill have no effect. But if the lower rated player actually wins the match then it will be counted. How much his rating will change depends on many factors - many of which have nothing to do with that particular match.

WTN is a complete mystery. It is unclear whether WTN considers all the games in a match or just who won each set. So a 6-0 6-0 might count the same as a 7-5, 7-6.

My 15-year-old junior ITF playing son beat this 1,100 ATP ranked Argentinian in two straight sets (7:5, 6:1) in M15 which he followed up with one more straight sets win (6:1, 6:2) win in the same Futures tournament. Things that don't sound right but are in our faces and that we cannot change are extremely frustrating. The powers in case that dictate us the rules of engagement while brushing our calls off, sweeping our concerns under the carpet and then they lie in our faces that their AI is working perfectly well 'cause they've checked it. They like us to identify an unclear issue as mystery so that they'd take no responsibility for their dificiencies and even aspirations.

If you think some areas have over rated players then have your son play there. He will gain rating points until their rating accurately reflects their skill level.

Just like if they had true coed tournaments I think the men playing against women in my area would improve their wtn rating. It would be a bonanza for rating improvement for a time but then it would equal out.

Yes, my son's currently playing in such an area but I'd not tag those players as "over rated"; I'd brand some players in some other locations underrated to their abilities instead. The problem that I have is that I cannot keep my 15-year-old junior ITF competitor in this "higher rated area" for long enough time due to a variety of circumstances. So, the boy plays both, the Futures and junior ITF; and, he does it in different nations on diferent continents every few weeks/months which is how those ATP, ITF, junior ITF opportunities come. I need to work and have my boy ranked in world juniors as well as in the ATP so that he gets his chances to sign into those better tournaments without having to beg for Wild Cards. Overall, I know down there that it is just a matter of time that my son reaches his deserved UTR 12 level although I am upset he needs to wait for that moment which should be here right at this very moment now.

Having WTN and UTR alongside those junior ITF and ATP ranking points is just not feasible, it's unethical and it opens doors to errors in judgment; I'd go as far as to say that all those kinds of players' ratings may be against laws in cases where for instance sponorships or college applications are turned down due to the UTR ratings.

 

[Moon Shooter]

I agree it is silly the way these tennis companies hide their algorithm. USTA does it because they say if certain people knew their rating they would start to cheat and throw games or something ridiculous like that. But that doesn't apply to WTN so they have no reason to hide the algorithm.

I doubt there is any discrimination going on against certain areas but that said we all saw a huge shift in US players ratings that clearly was not just a normal working of the WTN algorithm. I do not know if that shift happened in other areas. But I wouldn't be surprised if there were winners and loser regions from that "unusual intervention." Being transparent with the rating algo would just make the whole rating system more legitimate.

TennisBro what area are you from?

 

[TennisOTM]

My 15-year-old junior ITF playing son beat this 1,100 ATP ranked Argentinian in two straight sets (7:5, 6:1) in M15 which he followed up with one more straight sets win (6:1, 6:2) win in the same Futures tournament.

I think I found the player based on this data. That was a fantastic win against a UTR ~12.6 player who earned ATP points in 8 different tournaments last year, congrats to your son!

It's understandable to feel disappointed and angry that the UTR rating does not appear to be rewarding him with an appropriate increase for such an impressive win. However I do think it's important for any rating system to avoid over-reacting to one single result. You would want not that to happen in the other direction - if you son had a unusual bad day losing to a weak opponent you would not want his UTR to plummet just because of one bad result.

It does appear that this was by far his best result according to UTR. The win in the next round of that tournament was against a UTR 9 opponent. His rating currently includes 30 matches played in the last 7 months. He has had plenty of other opportunities to play UTR 11+ (mostly 11) opponents in those matches but has lost most of them, a 3-10 record by my count including some blowout losses to UTR 11 players.

You could argue many of the opponents who beat him have a deflated UTR as well, but those opponents come from many different countries and play a lot of matches with a big network of opponents. It seems unlikely that this group of players all happen to be deflated together, but maybe a more careful analysis would reveal something going on.

WTN is a disaster. According to your son's results page, he has beaten a WTN 6.4 and lost to a WTN 23.9. Many of his opponents are not even named: just says "player." What a mess. I really think they need to scrap the whole thing and start over from scratch.

 

[TennisBro]

@TennisOTM you might be on the right track there. You also seem to understand the situation pretty well there.

I think I found the player based on this data. That was a fantastic win against a UTR ~12.6 player who earned ATP points in 8 different tournaments last year, congrats to your son!

My boy beat a 6.5 WTN player ranked 1,105 in ATP, so I am now wondering if you have the right player there. But it truly was an unblievable straight sets win against a player with two pro coaches behind him and a couple broken tennis rackets in his angry second set that awared him with a code violation there. Note that my son travels alone and without any supporting staff. I wish I could give you my son's name but I fear repercussions. There are some bitter people in the system; take my word for it. They claim they use AI system a lot

TennisBro what area are you from?

Not NA or EU to say the least. But I now do send my son to some of those tournaments there.

 

[TennisBro]

Darn, @TennisOTM I stand corrected as you may have found my son's match but I hate to admit it for reasons I also hate.

Sorry for my inadequate consideration for your fine observation there; I am so crazily busy training on and off and then homeschooling my boy aside all the other crap i just cannot explain. I love coming here to TTW though. Sometimes, I feel it's the only place to vent your frustrations with the shortcomings of the system.