Why UTR is a deeply flawed system.

I don't think @MathGeek said the LC was going to unilaterally change the person's NTRP upward. All the LC was going to do was put the person in the next highest bracket.
That's the gist. The flaw in NTRP is that local USTA organizers are bound by ratings.

The difference is that UTR is just a rating body, they don't control what local organizers do. Thus, UTR ratings are just recommendations. Sandbagging players are not in control by virtue of dishonest ratings. Local tournament and league organizers are in control. They can place players according to their UTR ratings. OR NOT. Things that don't pass the smell test don't fly. The rating is not changed. It simply need not mandate how a local organizer places a given participant in their events. Other organizers are still free to honor a given sandbagger's UTR.

In USTA, the ultimate authority in competitive placement lies with the NTRP, which we've seen can be manipulated by sandbaggers. In UTR, the ultimate authority in competitive placement remains with the local organizer.
 
In my experience it has not been sandbaggers for the UTR tournaments just people that havent gotten an accurate UTR so the TD can adjust and get them good matches. Although the UTR tournaments I played in didnt have a "winner" per se either. You just played matches.
 
UTR is extremely inaccurate. Even Jim Courier and Lindsey Davenport have said in Tennis channel that it is just a number based on past record and isn't a accurate tool to assess who is playing better at the time of the match.
 
What number does provide an accurate tool to assess who is playing better at the time of a match? If not based on past performance, what would it be based upon?
 

schmke

Hall of Fame
I don't know that this says UTR is deeply flawed, but it seems to give little credit for losing competitive matches against higher rated players.

For example a player has 4 matches on their record going 1-3, in chronological order:

Played with a 7 vs a 9 and a 7 losing 6-3,6-3
Played with a 6 vs two 8's losing 6-4,6-4
Played with a 6 vs an 8 and a 7 losing 6-3,6-4
Played with a 7 vs two 8's winning 7-6,6-3

Current UTR is a 7 (6.68) which on the surface seems a bit low to my eye. Of course, just seeing whole numbers for the other players doesn't tell the whole picture.
 

MathGeek

Hall of Fame
If the match that drops off was boosting your rating then your UTR could drop.
I think most ratings systems would be better with weighted inclusion criteria rather than every match counting 100% or being completely excluded at 0%.

For example, rather than all included matches counting equally, older matches could gradually decline in how strongly they contribute to a current rating. Perhaps after a year, they only are weighted at 75% of more recent matches, and then gradually declining to 0% over the second year.

Similarly, perhaps match results considered to be from less reliable sources might simply be weighted less than being completely excluded.
 
I think most ratings systems would be better with weighted inclusion criteria rather than every match counting 100% or being completely excluded at 0%.

For example, rather than all included matches counting equally, older matches could gradually decline in how strongly they contribute to a current rating. Perhaps after a year, they only are weighted at 75% of more recent matches, and then gradually declining to 0% over the second year.

Similarly, perhaps match results considered to be from less reliable sources might simply be weighted less than being completely excluded.
UTR does that. Older matches count less. Shorter matches count less. And high rating spread matches count less (2.5 spread don't count at all)
 
And high rating spread matches count less (2.5 spread don't count at all)
Does this only count for people with verified ratings? It seems wrong to disregard a match with a wide spread when someone has a skewed rating due to lack of matches. How are they then supposed to get a more accurate rating?
 
Does this only count for people with verified ratings? It seems wrong to disregard a match with a wide spread when someone has a skewed rating due to lack of matches. How are they then supposed to get a more accurate rating?
Unless its over 2.5, the match still counts but just not as much as a closer match. If it is over 2.5 then it will ignore that match until you get some results to push you into the 2.5 range.
 
Unless its over 2.5, the match still counts but just not as much as a closer match. If it is over 2.5 then it will ignore that match until you get some results to push you into the 2.5 range.
Yeah I was talking about the over 2.5 ones. But at least it gets retroactively included once they get enough matches under their belt to put them within 2.5. That seems a bit more reasonable.
 
Top