Am I the only one who hates UTR?

#51
It encourages the idea that the score of a win matters. Think about that. That's like saying, Tom Brady's "rating" should be based on by how much he beat other teams. He has won 6 super bowls. But most of them are in a tight fashion, not in blowouts. I shouldn't be rewarded more or less for a close win vs a "decisive" one. The score doesn't tell the story of a match, and at the end of the day, winning is the most important thing, not killing your opponent.
I disagree. If you lose 0 and 0, the match probably wasn't very competitive, as opposed to a loss in two tiebreakers.

So obviously the number of games has to be factored into a system that has competitive matches as one of its main goals.
 
Last edited:
#52
UTR is completely worthless during the junior years, when you should be building your game for max potential after puberty and into young adulthood.

It becomes more relevant around 16-18 when game styles have become established, results become more important, and future potential can be pretty easily seen (for college coaches to assess).

It becomes utterly worthless again at the highest level on the ATP/WTA tour. No one cares.
 
#53
I disagree. If you lose 0 and 0, the match probably wasn't very competitive, as opposed to a loss in two tiebreakers.

So obviously the number of games has to be factored into a system that has competitive matches as one of its main goals.
You could lose 0 and 0 and the score in each game could be deuce 20 times where you have ad in your favor majorities of the times. Does UTR take that into consideration? It is like Federer had 23 break points against Djoker at 2015 USO and lost? Does UTR take that into consideration?
 
#54
You could lose 0 and 0 and the score in each game could be deuce 20 times where you have ad in your favor majorities of the times. Does UTR take that into consideration? It is like Federer had 23 break points against Djoker at 2015 USO and lost? Does UTR take that into consideration?
Of course UTR does not take that into consideration.

You can even win a match (as Nadal did against Thiem at last years US Open) and lose the UTR war (games 28-26 in favor of Thiem due to bageling Nadal in the first set). In any rating formula, it's just plain wrong to come out ahead of the person you lost to. And this happens quite frequently with juniors with super tie breaks in lieu of a third set.
 
#55
You could lose 0 and 0 and the score in each game could be deuce 20 times where you have ad in your favor majorities of the times. Does UTR take that into consideration? It is like Federer had 23 break points against Djoker at 2015 USO and lost? Does UTR take that into consideration?
Of course it doesn't, because it doesn't have access to that kind of data. And unless tournaments start recording point by point scores, game scores are as good as it gets.

But just think about what you are saying: "since we can't perfectly capture the match flow, let's throw the games out altogether, and only consider wins and losses!"

How isn't just going by match won/lost infinitely worse than considering the games won/lost?
 
#56
Of course UTR does not take that into consideration.

You can even win a match (as Nadal did against Thiem at last years US Open) and lose the UTR war (games 28-26 in favor of Thiem due to bageling Nadal in the first set). In any rating formula, it's just plain wrong to come out ahead of the person you lost to. And this happens quite frequently with juniors with super tie breaks in lieu of a third set.
If that's a problem, then tweak the formula so the overall result has more weight than games won/lost. It's not that difficult.

One would assume whatever formula they use gets adjusted on a regular basis for more accurate results. They have a huge data set to work with, and we don't.
 

jmnk

Hall of Fame
#57
Of course it doesn't, because it doesn't have access to that kind of data. And unless tournaments start recording point by point scores, game scores are as good as it gets.

But just think about what you are saying: "since we can't perfectly capture the match flow, let's throw the games out altogether, and only consider wins and losses!"

How isn't just going by match won/lost infinitely worse than considering the games won/lost?
because that's what tennis, or any other competition, is all about. To win. Not to get more yards, more time of possession, more shots on goal, longer drive in golf, etc, etc. You play the game to win. That's the only thing that matters. Which is why both UTR and NRTP are so flawed - they do not take into consideration the _only_ thing that matters.
 
#58
because that's what tennis, or any other competition, is all about. To win. Not to get more yards, more time of possession, more shots on goal, longer drive in golf, etc, etc. You play the game to win. That's the only thing that matters. Which is why both UTR and NRTP are so flawed - they do not take into consideration the _only_ thing that matters.
So, when you play rec tennis the only thing that matters to you is winning? You don't care if the matches are not competitive?

Wow congrats on that big win. You sure proved you're the greatest accountant/engineer/housewife to ever pick up a racquet in the Tri-County area

The goal of systems like the NTRP and UTR is to ensure competitive matches within similar ratings. Yes, they have flaws, but for the most part do their job just fine. The problem is some people give way too many much importance to some number on a computer
 
Last edited:
#59
The problem is some people give way too many much importance to some number on a computer
Completely agree with that comment.

I'm not sure if @jmnk is saying that UTR should only consider wins (maybe he is), but it seems like it should definitely be a factor. A lot of the systems in other sports that look consider a bazillion other metrics also use the fact that one team won and one team lost, and I think that tends to be the most important metric under consideration. All of the things discussed on this thread (scores, competitive vs blowout, home vs away, surface, etc) are interesting, but the win/loss has value too.
 
#60
If that's a problem, then tweak the formula so the overall result has more weight than games won/lost. It's not that difficult.

One would assume whatever formula they use gets adjusted on a regular basis for more accurate results. They have a huge data set to work with, and we don't.
Agreed games won should be taken into consideration but wins should have most weight, and games won in 3+ full sets should have less weight than set count. When a pro wins 3 out of 5, no one cares how many games each won or if one opponent had a bad set. It should be the same with collegians or juniors. If you win in 3 full sets, that should have the most weight along with the level of your opponent, but if the opponent won more games maybe the winner just gets a small boost, but under current algorithm, games won trumps sets won. UTR was originally created to determine competitiveness of matches which is totally different than predicting the winner of a match. Since UTR is used for recruiting and coaches want players who can win, more weight should be placed on wins while still counting games as a secondary factor. I believe 3rd sets played as tiebreaks count as 2 games so TBs dont have the weight of the first two sets which seems fair.
 
#61
I still dont really understand UTR, ive played tournaments in which I was sure to go up but went down.
I however am a big fan of it, so far based on UTR ratings all the players seemed to be somewhere around what they were rated at.

One thing to keep in mind, the UTR is the same for everyone, so if your daughter UTR can go up and down so can all other players. Algorithms cant be "unfair". Unless people start playing the system.
 
#62
One thing to keep in mind, the UTR is the same for everyone, so if your daughter UTR can go up and down so can all other players. Algorithms cant be "unfair". Unless people start playing the system.
UTR's algorithms may be fair or accurate within specific groupings, e.g. juniors girls 18s who play national tourneys, or top 25 team D3 players but are inaccurate across groups that seldom play each other. UTR is not just a measure of ability but also of opportunity. You will see players in the summer beat or have close matches with players UTR 1.5 ahead but the player who did better than expected does not move up much because maybe he/she only gets to play 4 matches like that out of his/her 30 counted matches (and also that 1.5 UTR gap gets less weight), and the rest of his/her matches are at a much lower UTR level. Same for the jr kids who cant afford to play nationals or ITFs or maybe even the D1 bench player who gets to play once or twice in lineup but mainly spends fall in low level flights. UTR would argue the under rated player should just totally dominate similarly ranked or lower ranked opponents, and his/her UTR would rise to the appropriate level, but it doesnt seem to work that way.
 

LGQ7

Hall of Fame
#63
Chess player here. UTR is the commercialization of the Elo rating system, without giving due credit to Elo.

I use the Elo rating system for my tennis Meetup. In applying the Elo equation I come across a phenomenon that I correct and I call it "The Sylvester Rule".

I count to 21. 20-20 is a draw. I use the initial score to assign a new Elo rating. Makes sense. After that I only count win/loss/draw.

In assigning new players. Let's say a new player beats the best known player (2000) by 21-15, but he beats the average player (1500) by 21-10. Which do I count? The best? The average?

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

Best = 2058
Worst = 1629

I go with the best and not the average. The new person's rating is 2058 + a win over 1500 = 2059.

I call it the Sylvester Rule. It comes from the Sylvester and Tweety cartoons. Sylvester has a son. His son runs into a kangaroo who he thought was a giant mouse. Scarred out of his wits he runs into his father, Sylvester. Sylvester asks his son how big is the mouse. His son stretched out his arm as wide as possible and said, "This big." Sylvester measure the maximal wingspan of his son and it is very little to him. He laughs at the length he measures and said, "Oh, this big?"

That's how I came up with the rule and called it "The Sylvester Rule."

Somebody said before that only win/loss are meaningful measurement in tennis. That is correct. Win in games should only be used as the initial rating to get "more data points".

https://tt.tennis-warehouse.com/ind...appeal-of-tennis-ratings.616313/post-12302379
 
Last edited:

LGQ7

Hall of Fame
#64
Chess player here. UTR is the commercialization of the Elo rating system, without giving due credit to Elo.

I use the Elo rating system for my tennis Meetup. In applying the Elo equation I come across a phenomenon that I correct and I call it "The Sylvester Rule".

I count to 21. 20-20 is a draw. I use the initial score to assign a new Elo rating. Makes sense. After that I only count win/loss/draw.

In assigning new players. Let's say a new player beats the best known player (2000) by 21-15, but he beats the average player (1500) by 21-10. Which do I count? The best? The average?

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

Best = 2058
Worst = 1629

I go with the best and not the average. The new person's rating is 2058 + a win over 1500 = 2059.

I call it the Sylvester Rule. It comes from the Sylvester and Tweety cartoons. Sylvester has a son. His son runs into a kangaroo who he thought was a giant mouse. Scarred out of his wits he runs into his father, Sylvester. Sylvester asks his son how big is the mouse. His son stretched out his arm as wide as possible and said, "This big." Sylvester measure the maximal wingspan of his son and it is very little to him. He laughs at the length he measures and said, "Oh, this big?"

That's how I came up with the rule and called it "The Sylvester Rule."

Somebody said before that only win/loss are meaningful measurement in tennis. That is correct. Win in games should only be used as the initial rating to get "more data points".

https://tt.tennis-warehouse.com/ind...appeal-of-tennis-ratings.616313/post-12302379
If the new player beat the best player but lost to an average player, then the average Elo rating will make sense for his initial rating. That's the application of "The Sylvester Rule".
 

ClarkC

Hall of Fame
#66
In the Elo ratings in chess, you have wins and losses. No one says "You put up a better fight than expected against that higher-ranked player; you made it to move 55 in the end game, so you get more credit for losing than he does for winning." The whole controversy here is the counting of games won, which can outweigh who actually won the match.
 
#67
Does ELO count the games in chess that make up a final, like when Fischer beat Spassky 5 to 4? So Spassky "wins" anyway if he was expected to only get one draw and no wins?
 
#68
UTR is basically a money-making scam. The original message was that they were going to grow the game, level the playing field and make tennis more accessible for everyone. UTR has reversed course. The social function and self-reporting function were supposed to be released in the fall of 2018 - still waiting. They've also abandoned the players report, where you could screen for players of the same ability in a similar geography. The only "social" function is to search for paid events. They claim that subscribers are supposed to get a discount on UTR events. My daughter has played in over 75 UTR events and I've never received a discount.

UTR events are good for 10U kids who don't want to go through the orange/green ball progression and for kids who always get beat in the first round of a USTA tournament. For everyone else, cross-play is counterproductive. My daughter got matched up with older boys and girls, who are naturally bigger and stronger. It forced her to play defensively, which she would never do against girls her own age. She should have been playing girls in her age range and use her current and future strengths - play offense and create opportunities.

Their customer service is very deceptive and/or ignorant.

As for ratings, if someone starting out as a 1.5 beats a couple players rated 5.0, it doesn't count towards either players' UTR - dumb. I hope the rating is just a starting point in the recruiting process. It would be a shame if college coaches depended on such a flawed system.
 
#69
UTR is basically a money-making scam.
Not a scam, but certainly a business. If it was truly about a better rating system they could have worked with the USTA to improve NTRP. But look who is the driving force behind UTR...TC execs, business men, and a guy with an algorrythm. They are working and promoting hard to peal away players for UTR only tournaments at some point, but I see a lot of issues moving forward.
 
#70
I dont find it a scam. Here in Illinois the search function works great to find adult tournaments. There is one almost every other weekend. They usually are rated 6 and below, 8 and below and open, from what I have seems the ratings seem to work.
I am not a paying member, but I think most tournaments have a $10 discount for them.

I am a fan of USTA tournaments also, but their search engine is terrible, and having to buy the membership always bothered me.
 
#71
An example of how UTR is skewed by lack of opportunity and crossplay: the D3 national indoors was completed this weekend. There were guys who were 5 stars as juniors playing in the lineup for the top teams. However, over the last 2 1/2 years, the UTRs of some players have dropped more than 1.0. Looking at UTR, a high school player may think he could play towards the top of the lineup of a top 8 D3 team with a UTR 12.0. What that HS recruit does not realize is some of those players were UTR 13s when recruited in their jr/sr year. In spite of the UTR drop, these guys have improved their game-especially in dubs. Most are underranked at least .5 because they play other underranked D3 players. However, they are enjoying their college careers so maybe the message should be once players are recruited to ignore their UTRs and just focus on continual improvement. UTR goes through algorithm changes, and I think most college players-except those that regularly play in the lineup of top 40 teams-have seen their UTR drop .4-.5 across the years even when winning the same % of matches.
 
#72
My singles UTR continues to plummet, despite zero activity. I think there is something wrong with the algorithm - my UTR shouldn't change over time just because my opponents from more than a year ago start improving or declining.

My UTR has gone down by 2 UTR units in the past several months, even though my last match played was well over a year ago. If USTA ratings worked that way, you'd have people cheating by simply taking a season off to get themselves bumped down.
 
#73
Not a scam, but certainly a business. If it was truly about a better rating system they could have worked with the USTA to improve NTRP. But look who is the driving force behind UTR...TC execs, business men, and a guy with an algorrythm. They are working and promoting hard to peal away players for UTR only tournaments at some point, but I see a lot of issues moving forward.
If a product or service is misrepresented, it's a scam - not a legitimate business.
 
#74
My singles UTR continues to plummet, despite zero activity. I think there is something wrong with the algorithm - my UTR shouldn't change over time just because my opponents from more than a year ago start improving or declining.

My UTR has gone down by 2 UTR units in the past several months, even though my last match played was well over a year ago. If USTA ratings worked that way, you'd have people cheating by simply taking a season off to get themselves bumped down.
Not really, because UTR does not reward sandbaggers. USTA cheaters like to play down a level so they can go to Nationals or collect a plastic tourney trophy whereas UTR groups you with people closest to you in rating with the goal of improving your UTR rating and then being grouped with higher UTR rated players in the next tourney.


Sent from my iPhone using Tapatalk
 
Top