[BostonCard]

Thanks for this ColoradoTree.

I agree that some sort of test of how well UTR works in practice in predicting the outcome of matches.  It seems like this is a problem tailor made for an Elo approach (which UTR doesn't seem to be, but may use some aspects of).  From the wiki:

https://en.wikipedia.org/wiki/Universal_Tennis_Rating  

In contrast, designing draws or flights with UTR consistently produces many more matches between athletes of roughly equal ability. Two players whose UTRs fall within 1.0 of each other will have a competitive match two to three times as often as those with ratings more than 1.0 apart. Furthermore, research has shown that when players' UTRs differ by more than 1.0 point, the lower-rated athlete will upset his or her higher-rated opponent only 3 percent of the time.

BC

[BostonCard]

Thanks for this ColoradoTree.

I agree that some sort of test of how well UTR works in practice in predicting the outcome of matches.  It seems like this is a problem tailor made for an Elo approach (which UTR doesn't seem to be, but may use some aspects of).  From the wiki:

https://en.wikipedia.org/wiki/Universal_Tennis_Rating  

In contrast, designing draws or flights with UTR consistently produces many more matches between athletes of roughly equal ability. Two players whose UTRs fall within 1.0 of each other will have a competitive match two to three times as often as those with ratings more than 1.0 apart. Furthermore, research has shown that when players' UTRs differ by more than 1.0 point, the lower-rated athlete will upset his or her higher-rated opponent only 3 percent of the time.

BC