ITF ups the ante in the "global" ratings game

[schmke]

I just wrote on my blog about the ITF announcing the World Tennis Number (WTN) which seems to be their entry in the ratings game to go head to head with UTR.

To be fair, the ITF created the International Tennis Number well over 10 years ago so pre-dated UTR, but had clearly not invested in having it adopted and used the way UTR has. With the support of the major tennis associations, the WTN seems to be a replacement and modernizing of ITN and an effort to have it be calculated if not adopted by players from many countries.

No, I don't expect NTRP to go away any time soon. But the USTA is one of the supporting national associations and so I'm guessing will enable the ITF to calculate WTNs or perhaps even calculate it themselves for players participating in USTA leagues and tournaments.

It will be interesting to see what happens!

 

[BallBag]

Its identical to UTR with no gender and age separation and has separate singles and doubles ratings. One key difference is that they are saying the calculation will be open.

USTA could even roll WTN into the tennislink overhaul. I would really like to see dynamic brackets for league divisions instead of NTRP.

 

[travlerajm]

I like the idea of open calculation. Keeps it honest.

 

[schmke]

Yeah, let's see how "open" it really is. It really should be completely open and transparent allowing any federation/association to implement it themselves. My guess is it will be open in that how it works will be described, but some details required to actually replicate the calculations may not be available, in part to "encourage" the use of authorized ratings sites (ClubSpark is mentioned in the PR).

 

[BallBag]

@jcgatennismom
I found these presentations from ITF Academy about WTN. Some interesting highlights:

The ITF is developing a Tennis Open Data Standard (TODS) for the sharing of match data. Meaning no more web scraping and/or individual data parsing methods for each data source. No interns manually entering data either. They didn’t say if anyone would be able to download the data. Could be restricted to authorized national associations and vendors.

ITF owned World Tennis ID. The ITF will assign an ID and sign-on to each player to avoid duplicating players. It will also create infrastructure for a single sign-on (like Facebook sign-on) to keep track of players.

They are using the Glicko-2 method for rating calculation, which shares a lot of features with UTR so I’m wondering if UTR is also using Glicko or Glicko-2. This means that even if they do not completely share the math, it should be fairly straight forward to get to it. There is an assigned system constant , τ , and then some initialization parameter. Also, Glicko-2 uses a scale of 0 to 1 for match performance, I don’t know how they will convert match scores to that 0 to 1 scale.

The IFT is also developing software for tournament administration (registration, scheduling and draw). I doubt the USTA will cede that to the ITF but I’m sure there are national organizations that don’t even have a tennislink type platform so I’m guessing it’s for those organizations.

Overall, seems like they are duplicating what UTR is doing on a global level. This is really validating what UTR is all about but it’s also fragmenting the tennis landscape. They are saying that they should start launching by the end of this year so we’ll see what this looks like.

 

[ChaelAZ]

Interesting the way the title reads on this UTR advertisement article, that Fed and Djo support UTR but nothing is ever mentioned to that effect. They also imply other ratings are failed, which is not factual. Flawed, but certainly not a failure.

Also interesting is Mark's comments about the business model of UTR. As a HUGE ventre capitalist, while I appreciate the freemium model, it is about the upsell and profits. Venture capitalists are not doing philanthropy. Worth a read.

Djokovic & Federer buy into new tennis ratings system. Can Universal Tennis Rating succeed where others are failing?
https://**********.net/djokovic-fed...nnis-rating-succeed-where-others-are-failing/

"The UTR could also be of use in countries which do not have the same kind of resources as the Grand Slam nations. In South Africa, for example, where the federation has almost no funds to help send junior players overseas, staging more UTR events at home would allow them to judge the standard of their players against those in other nations, without the cost, both financial and environmental, of flying them around the world.

Though the British governing body of tennis, the LTA, sends results from its players to be included in the UTR, so far it has been reluctant to join forces. In the past 30 years, the British ratings system has changed a number of times and the LTA has just announced it is working with the International Tennis Federation on another new global ratings system called the World Tennis Number. You an learn more here.

Leschly hopes to work with the LTA in future. “At this point, we continue to rate their players – we’re offering a free service there – but I think there’s so much more we could do for UK tennis and I hope we have the opportunity to do that. We’re here, we’re a tool. There’s nothing we’ve ever done or want to do to undermine federations. Federations are fundamental to support tennis in the various nations. We’re not trying to lock anyone out, we’re not asking for exclusivity. This is a free market. We want to democratize tennis.

“If people want to work with us, great. Almost all of our technology is free and we just want to drive huge adoption. Our business model long term is: Hey, if people want to use this platform, there’s benefit for them. Maybe somewhere along the line we can up-sell them something of incremental benefit. It’s a freemium model, there’s no Trojan horse, we’re regulatory compliant. All of this stuff is about opt-ins.”

So WTN seems to align with the USTA/ITF and International govenring associations to support their efforts and is controlled within them, not as an outsourced service. I personally think that is a better model.

 

[jcgatennismom]

I just wrote on my blog about the ITF announcing the World Tennis Number (WTN) which seems to be their entry in the ratings game to go head to head with UTR.

To be fair, the ITF created the International Tennis Number well over 10 years ago so pre-dated UTR, but had clearly not invested in having it adopted and used the way UTR has. With the support of the major tennis associations, the WTN seems to be a replacement and modernizing of ITN and an effort to have it be calculated if not adopted by players from many countries.

No, I don't expect NTRP to go away any time soon. But the USTA is one of the supporting national associations and so I'm guessing will enable the ITF to calculate WTNs or perhaps even calculate it themselves for players participating in USTA leagues and tournaments.

It will be interesting to see what happens!

There is a similar thread on the college forum https://tt.tennis-warehouse.com/ind...r-2020-college-coaches-use-wtn-or-utr.647742/
What is your blog?

 

[schmke]

There is a similar thread on the college forum https://tt.tennis-warehouse.com/ind...r-2020-college-coaches-use-wtn-or-utr.647742/
What is your blog?

I'm not allowed to promote on Talk Tennis by posting links (others can post to my site though), but Google Schmidt Computer Ratings and you should find it.

 

[LGQ7]

Ho hum. Old news to chess players. The Elo rating system was instituted in 1960, that's about 60 years ago. UTR is an adoption of Elo.

Right here, if you want to use it for your tennis club.

http://www.3dkingdoms.com/chess/elo.htm

 

[Shankadelic]

There is a similar thread on the college forum https://tt.tennis-warehouse.com/ind...r-2020-college-coaches-use-wtn-or-utr.647742/
What is your blog?

http://computerratings.blogspot.com/2019/07/itf-announces-new-world-tennis-number.html

 

[LGQ7]

Ho hum. Old news to chess players. The Elo rating system was instituted in 1960, that's about 60 years ago. UTR is an adoption of Elo.

Right here, if you want to use it for your tennis club.

http://www.3dkingdoms.com/chess/elo.htm

Let me give you an idea of how long ago that was. One of the first portable electronic calculator is a The Faber-Castell TR3 introduced in 1973. It had a slide rule back up (the predecessor of electronic calculators).

I know you guys don't know what a slide rule is for. It was foretold in a prophecy.

 

[LGQ7]

In 1960, this was the calculator.

 

[LGQ7]

Let me give you an idea of how long ago that was. One of the first portable electronic calculator is a The Faber-Castell TR3 introduced in 1973. It had a slide rule back up (the predecessor of electronic calculators).

I know you guys don't know what a slide rule is for. It was foretold in a prophecy.

Wonderful World was recorded in 1959 and released in 1960.

Forensics linguistics archaeology. A slide rule is the calculator of that time.

 

[BallBag]

One difference between WTN and UTR will be that WTN will not recalculate your rating unless you have new data, Glicko-2 doesn't support that.

 

[BallBag]

Some new info on WTN. Scale will be 40 to 1 with one decimal resolution, 40 being complete beginner. Unrelated to that, the LTA requires a certain playing level for their coaching cert. Does the USTA do that?

 

[sureshs]

Some new info on WTN. Scale will be 40 to 1 with one decimal resolution, 40 being complete beginner. Unrelated to that, the LTA requires a certain playing level for their coaching cert. Does the USTA do that?

USTA does not do certs. It is done by the USPTA or USPTR. USTA has recently decided to exert its power and require coaching bodies to become accredited, and till now only the USPTA has done so.

For the USPTA or USPTR, I have heard 4.0, but I think it is based on assessment of your play during the testing part of the exam, rather than a league record. You will be asked to hit every type of shot and evaluated on that.

 

[BallBag]

So it's been a year and no updates from the ITF.
One additional bit of info I found is that it will be based on sets instead of games. I guess it takes the pressure off juniors to win every game but it also means that it will next to useless for rec players.

 

[schmke]

So it's been a year and no updates from the ITF.
One additional bit of info I found is that it will be based on sets instead of games. I guess it takes the pressure off juniors to win every game but it also means that it will next to useless for rec players.

Why do you say it will be next to useless for rec players?

While I agree finer grained data it better than coarse grained data, I don't think it is that much different. There are ratings algorithms that look at just who wins/loses ignoring score, and they've proven to be pretty accurate as unless a player is only ever playing players better than them, eventually they win a match and the algorithm does its job. Similarly, using just sets offers finer grained detail, and given enough matches, it works itself out.

If your point is that rec players play few enough matches that the finger grained detail of games is needed, I think you do have a point, I think games allows for a more accurate rating with fewer matches played, but I wouldn't call using just sets next to useless.

 

[BallBag]

Why do you say it will be next to useless for rec players?

While I agree finer grained data it better than coarse grained data, I don't think it is that much different. There are ratings algorithms that look at just who wins/loses ignoring score, and they've proven to be pretty accurate as unless a player is only ever playing players better than them, eventually they win a match and the algorithm does its job. Similarly, using just sets offers finer grained detail, and given enough matches, it works itself out.

If your point is that rec players play few enough matches that the finger grained detail of games is needed, I think you do have a point, I think games allows for a more accurate rating with fewer matches played, but I wouldn't call using just sets next to useless.

If you treat each score as a measurement sample than the full scale resolution of a WTN match is between 0 and 3 with a resolution of 1/4. 0 for staight set lose, 3 for staight set win and then something in between for split sets each way. A straight set UTR match with 1 game resolution is something between 0 and 12 so 1/13. So each UTR match would provide 3 times as much data as a WTN match. Considering UTR is not very accurate for rec players that play mostly dubs, WTN would be even less accurate. One solution could be for WTN to count matches extending 3 years to compansate where as UTR only counts 1 year.
I agree that for active singles players, the rating would have enough data to be accurate. Most rec players do not play that much singles and they do not have many matches on record so WTN could not hope to rate them accurately.

 

[Doan]

If you treat each score as a measurement sample than the full scale resolution of a WTN match is between 0 and 3 with a resolution of 1/4. 0 for staight set lose, 3 for staight set win and then something in between for split sets each way. A straight set UTR match with 1 game resolution is something between 0 and 12 so 1/13. So each UTR match would provide 3 times as much data as a WTN match. Considering UTR is not very accurate for rec players that play mostly dubs, WTN would be even less accurate. One solution could be for WTN to count matches extending 3 years to compansate where as UTR only counts 1 year.
I agree that for active singles players, the rating would have enough data to be accurate. Most rec players do not play that much singles and they do not have many matches on record so WTN could not hope to rate them accurately.

Looks like WTN is using 4 years worth of data. Maybe they will use this as paper on "Hierarchical Glicko-2" to help with the issue when people in different groups do not play each other.

https://www.imperial.ac.uk/media/imperial-college/faculty-of-engineering/computing/public/1920-ug-projects/distinguished-projects/Bringing-the-World-Tennis-Number-to-Life-(International-Tennis-Federation).pdf

 

[BallBag]

I found this WTN Presentation on the ITA website.

Key Points:

• WTN will include results from USTA, ITF and some ITA (no results from UTR tournaments)
• The scaling was done with 1 being highest to support some sanctioning bodies using WTN for rankings
• 4 years of data will be used for initial rating and then phased out in favor of more recent results
• Results from higher class events will be weighed more towards your rating. So results from a L1 tournament will influence your rating more than a L3. (This is a global thing and USTA has the option to weigh all tournament equally)
• Match format weights:

• Juniors will get profiles first
• USTA is considering if WTN can replace NTRP but there are no immediate plans
• No timeframe for rollout

 

[Wild Card]

Britains LTA has introduced a WTN separate ratings for Singles and Doubles.
- Results from all nations supporting the World Tennis Number will be used by the algorithm to count towards a player’s Number.
- allegedly results from some college tennis matches will be included
Wonder how the WTN will impact UTR? And how soon other countries will begin the roll out?

ITF World Tennis Number

The ITF World Tennis Number provides a global standard for all tennis players, including doubles.

www.lta.org.uk

 

[Harry_Wild]

If USTA and other worldwide tennis organizations adopts the WTN ranking system, UTR and NTRP systems is finished, obsolete! We basically have one system worldwide. Of course we already know where the top 500 ATP and WTA Tour players stand now but we will know where everyone else stands if the play competitive tennis matches that are in some sort of association that reports results into computer tennis database. ITF tour level players need to get their ranking up there till entered the ITF Transition level bottom tier tournaments. Anybody know want the current cut of is now?

 

[PBODY99]

I still have my 10" deci log log metal "slip stick " from 1966

 

[Tennis Sam]

Are there any lists available anywhere that show WTN's yet? I've seen a few instances of individual player profiles at the ITF website, but no lists. This has been a LONG time in the making. I hope it lives up to the hype.

 

[BallBag]

The USTA was going to roll out WTN for junior tournaments first. I don't think it's been implemented there yet so adults very far behind.

 

[Moon Shooter]

She does say that the algorithm will take into account "scores" of the sets. So I take that to mean it will be sensitive not just to who won the set but it will also consider the number of games won! See the video at 42:26.

I am not sure what she is saying here:

Is she saying that people can run events *and enter the scores* from matches outside of official USTA events? Or is she just saying people outside of USTA can use the ratings to run their own events? If it is the former this will be a great system. If it is the latter I am not sure how USTA would charge for that information as presumably anyone in USTA should be able to see other peoples ratings.

Also it seems that the European Union is so draconian it has laws preventing people from sharing the score of a tennis match!