DavidC wrote:

---

I'm not familiar with the Universal Tennis Ratings but they sound like a more scientific version of my attempts to rapidly assess the quality of players based on how they have fared against various opponents in junior and pro tennis. I wonder how they cope with sudden step changes in capability, which is all part of the junior learning curve.

SF

(8) Kseniya Yersh (BLR) v Nell Miller

Doubles F

(3) La Rocca / Miller (ITA/GBR) v Domnina / Pepelyaeva (RUS)

---

Obviously I don't know the exact algorithm, but I've had some dealings with UTR since launching my website.

They use the last 30 matches, providing they were played within a year of the present day. So, matches played February and January of last year no longer contribute towards the UTR even if they fall within a player's last 30 competitive matches. So if a player does see a huge increase in form or ability, then the rating will, eventually, grow to reflect that, providing the player continues to secure good results against good players on a regular basis. 

It's kind of like chess: If I'm rated 1000, my rating goes up more if I beat a 1200 rated player than it does if I beat a 900 rated player. In tennis, the scoreline also affects the UTR, so beating an opponent 6-0 6-0 boosts your UTR more than it would if you beat the same opponent 6-4 6-4 or if you dropped a set.

They're trying to secure results from LTA, Tennis Europe, ITF and other national tennis federations in order to make the Ratings more accurate for every player.

So, in essence, David, a more scientific way of what you were already doing