UTR Flex League Match Blog

[TennisOTM]

I’m guessing that the UTR algo can’t handle him because he has played against too wide a range of players.

The algo has a stupid piece of code that tells it to ignore matches where there is >2.0 UTR units level gap between opponents, unless the lower rated player wins.

So in this case, the algo is confused about whether it should count his 1 game won result vs. a UTR 9.xx and ignore his wins against 5.xx opponents. So it hedges and discounts his reliability.

To illustrate how poorly considered this piece of code is, consider the situation with 8.0 mixed, where a typical match might pit a UTR 8/3 combo against a 6/5.

The majority of these mixed matches don’t count toward the 8’s rating, but in the 30-40% of the time that the 8/3 loses to the 6/5, it counts. Thus biasing the ratings of all 4 players.

You can tell which match results are counted by the tennis ball icon next to the score (says "used for UTR calculation" when you hover over it), and that icon is normally missing for the >2.0 gap matches like you describe. Yet for this guy, the icon for is there for his wins over UTR 4.XX and even 3.XX opponents, and it's also there for his losses against UTR 9.XX opponents. How is that possible?? He's like some kind of Schrödinger's cat UTR 5 and 7 at the same time.

 

[travlerajm]

You can tell which match results are counted by the tennis ball icon next to the score (says "used for UTR calculation" when you hover over it), and that icon is normally missing for the >2.0 gap matches like you describe. Yet for this guy, the icon for is there for his wins over UTR 4.XX and even 3.XX opponents, and it's also there for his losses against UTR 9.XX opponents. How is that possible?? He's like some kind of Schrödinger's cat UTR 5 and 7 at the same time.

My wag is that he will wake up tomorrow with his UTR downgraded from 7.x to 5.x (and reliability increased to at least 50%), and my UTR will be unaffected, just used to help calculate his. We shall see.

 

[schmke]

You can tell which match results are counted by the tennis ball icon next to the score (says "used for UTR calculation" when you hover over it), and that icon is normally missing for the >2.0 gap matches like you describe. Yet for this guy, the icon for is there for his wins over UTR 4.XX and even 3.XX opponents, and it's also there for his losses against UTR 9.XX opponents. How is that possible?? He's like some kind of Schrödinger's cat UTR 5 and 7 at the same time.

I think UTR has an oscillation issue where they do an iteration where a 5 beats several 4s and those wins are counted, but the result of that is the 5 moves to a 6 or 7, and then in the next iteration those aren't counted due to >2.0 difference and the player comes back down. The rating they show for the player just depends on which iteration they stopped after.

 

[travlerajm]

I think UTR has an oscillation issue where they do an iteration where a 5 beats several 4s and those wins are counted, but the result of that is the 5 moves to a 6 or 7, and then in the next iteration those aren't counted due to >2.0 difference and the player comes back down. The rating they show for the player just depends on which iteration they stopped after.

Yet another reason that the >2.0 rule is messing with the accuracy.

The defense of the >2.0 rule is that games lost by the better player in these cases aren’t informative because the better player might be coasting and ceding freebie games.

My counter argument is that if a UTR 12 is giving any free games to a UTR 8, that maybe he deserves to get dinged for it.

 

[schmke]

Yet another reason that the >2.0 rule is messing with the accuracy.

The defense of the >2.0 rule is that games lost by the better player in these cases aren’t informative because the better player might be coasting and ceding freebie games.

My counter argument is that if a UTR 12 is giving any free games to a UTR 8, that maybe he deserves to get dinged for it.

I think they have this clause so that the UTR 12 doesn't look at a draw and see they'll play an 8 and choose not to play because they know giving up a game will ding them. Yes, this happens with juniors who live by (and get recruited based on) their UTR.

 

[travlerajm]

I think they have this clause so that the UTR 12 doesn't look at a draw and see they'll play an 8 and choose not to play because they know giving up a game will ding them. Yes, this happens with juniors who live by (and get recruited based on) their UTR.

If a UTR 12 is really a UTR 12 and not a soft poser, he shouldn’t have much trouble with the challenge of double-bageling a UTR 8.

Even if we go along with assumption that a UTR 8 can take a game off a UTR 12, and it might even be expected, then it seems like the algo could easily find a way to handle it (by adjusting the expected score) rather than excluding the match altogether.

 

[TennisBro]

I think they have this clause so that the UTR 12 doesn't look at a draw and see they'll play an 8 and choose not to play because they know giving up a game will ding them. Yes, this happens with juniors who live by (and get recruited based on) their UTR.

Yes, juniors look at their UTRs for the recognition. Sponsors and college recruiters may look at the ratings but the ITF rankings take it symbolically. You won't get seeded in the tournament based on your UTR rating, so why pay attention to it. The rating's useless basically.

 

[Dags]

I don't like to wish ill on anyone, but let's hope this particular opponent gets a pasting from all participants, and the algo sorts out his correct position. Winning two points in a set isn't fun for anyone involved, no matter how nice they are about it.

Fingers crossed none of the others see this result and start ducking our TTW hero like they did previously!

 

[travlerajm]

I don't like to wish ill on anyone, but let's hope this particular opponent gets a pasting from all participants, and the algo sorts out his correct position. Winning two points in a set isn't fun for anyone involved, no matter how nice they are about it.

Fingers crossed none of the others see this result and start ducking our TTW hero like they did previously!

UTR ratings skipped the usual daily update this morning. It’s almost like this guy has a force field around him that immunizes him against ratings.

 

[Dags]

UTR ratings skipped the usual daily update this morning. It’s almost like this guy has a force field around him that immunizes him against ratings.

Maybe his results are crashing the system?

 

[FuzzyYellowBalls]

I think my prediction of 100% UTR 5 will eventually happen, I have a brain for UTR lol. He's a 64% 6 now. Suppose it depends if he is now a 6.2 or a 6.8 for example.

 

[travlerajm]

I think my prediction of 100% UTR 5 will eventually happen, I have a brain for UTR lol. He's a 64% 6 now. Suppose it depends if he is now a 6.2 or a 6.8 for example.

6.8 strokes with 6.2 shot tolerance awareness. I think the 1 game he won against the UTR 9.x is suspending him above 6.

 

[FuzzyYellowBalls]

6.8 strokes with 6.2 shot tolerance awareness. I think the 1 game he won against the UTR 9.x is suspending him above 6.

Hahaha, nice.

 

[travlerajm]

As an aside, I’m thinking I might write to Quinn to make the case for dropping the >2.0 gap rule in their algo.

About 18 months ago, I discovered a serious system wide bug in the UTR algo for calculating doubles ratings. I wouldn’t have caught it if I wasn’t a power subscriber at the time, but the performance charts in my profile made the bug obvious and easy to figure out what caused it. It was an error in the code that omitted accounting for partner strength when calculating match ratings. The code was correct for calculating opponent strength for the performance charts, but not for the actual rating. The fact that I was playing exclusively 8.0 mixed at the time made it obvious.

I passed the bug on to quinn, who forwarded it to the UTR data team. The repeated reply (after several attempts at logical arguments) from the data team was that they checked it and it’s working perfectly. So I gave up.

However, 6 months later, the bug got fixed. And doubles ratings have been properly accounting for partner strength since they fixed it about 12 months ago.

So the mechanism for fixing the algo from user feedback exists, but whoever is in charge of the data team has a cya denial approach.

 

[TennisOTM]

I guess it's possible that dropping the >2.0 gap rule would make a sudden improvement, but I'm guessing it's much more complicated than that.

First, I think they give more weight to match results between opponents with a lower gap. So it's not like all matches with <2.0 gap are weighted the same and >2.0 weigh nothing with a hard cutoff. I'm thinking the weight gradually decreases to zero at 2.0. So a change to this part of their algorithm would not be a simple rule switch but would mean changing their weighting formula, which might have difficult to predict consequences.

Second, I think the larger issue is probably with the iterations that @schmke has described. Because everyone's rating affects all of their opponents' ratings and vice versa, there's a circularity that can be difficult to handle mathematically. Re-running calculations iteratively is one way to handle it, where each iteration will get the ratings closer to converging to an optimum. Most active players probably don't change much after each iteration, so UTR does not bother running many of them every single day. But then there are the weird edge cases you've identified who probably do need at least several more iterations to settle down to something reasonable.

 

[travlerajm]

I guess it's possible that dropping the >2.0 gap rule would make a sudden improvement, but I'm guessing it's much more complicated than that.

First, I think they give more weight to match results between opponents with a lower gap. So it's not like all matches with <2.0 gap are weighted the same and >2.0 weigh nothing with a hard cutoff. I'm thinking the weight gradually decreases to zero at 2.0. So a change to this part of their algorithm would not be a simple rule switch but would mean changing their weighting formula, which might have difficult to predict consequences.

Second, I think the larger issue is probably with the iterations that @schmke has described. Because everyone's rating affects all of their opponents' ratings and vice versa, there's a circularity that can be difficult to handle mathematically. Re-running calculations iteratively is one way to handle it, where each iteration will get the ratings closer to converging to an optimum. Most active players probably don't change much after each iteration, so UTR does not bother running many of them every single day. But then there are the weird edge cases you've identified who probably do need at least several more iterations to settle down to something reasonable.

The gyrating rating oscillations and hard cap on the weighting aren’t even the dumbest parts of the >2.0 rule.

The poorest considered part of the rule is the clause that says >2.0 gap matches count only if the lower rated player pulls the upset. By counting these upsets, but not counting the almost upsets or the expected spankings in this gap range, this biases all the ratings in the system in multiple ways.

The introduction of this is backward thinking. They see that upsets in the >2.0 gap range correlate with level better than routs in this gap range. Therefore they should count!

 

[schmke]

The gyrating rating oscillations and hard cap on the weighting aren’t even the dumbest parts of the >2.0 rule.

I don't think the gyrating rating oscillations are tied to the >2.0 rule. I've played few if any opponents farther than 2.0 away from me, but my UTR regularly oscillates 0.5 or more. In fact, yesterday I was a 6.22 and today I'm a 6.75, both 100% reliable with 33 matches played in the past year. No new matches recorded for me or any of my partners.

Now, perhaps they got around to dropping a result from 1/29 a year ago that was a loss (but not a terrible one), but the 6.22 was initially published on yesterday (1/31) and the 6.75 today (2/1), so I would have thought any dropping of results a year+ old would have affected both.

 

[travlerajm]

I don't think the gyrating rating oscillations are tied to the >2.0 rule. I've played few if any opponents farther than 2.0 away from me, but my UTR regularly oscillates 0.5 or more. In fact, yesterday I was a 6.22 and today I'm a 6.75, both 100% reliable with 33 matches played in the past year. No new matches recorded for me or any of my partners.

Now, perhaps they got around to dropping a result from 1/29 a year ago that was a loss (but not a terrible one), but the 6.22 was initially published on yesterday (1/31) and the 6.75 today (2/1), so I would have thought any dropping of results a year+ old would have affected both.

The gyrating ratings are worse in mixed. At one point, my rating was oscillating back and forth daily between “100% reliable” UTR 7.x and “50% reliable” UTR 9.x despite me not having played a match in months.

(I think) it’s because half my matches (the blowout wins) were being ignored by the algo on Tuesdays because my rating was >2.0 better than my opponent mixed rating (calculated as sum of opponent UTR minus partner UTR). Then on Wednesdays, my UTR was low enough that all the matches got counted on the next iteration, causing my rating to spike up above 9. Which in turn caused my Thursday algo rating to ignore the 50% of my >2.0 gap matches again.

The imbalanced nature of combo rating 7.0, 8.0, and 9.0 mixed means that the >2.0 gap rule applies to roughly half the matches with combo rated teams, destabilizing the algo.

 

[TennisOTM]

The imbalanced nature of combo rating 7.0, 8.0, and 9.0 mixed means that the >2.0 gap rule applies to roughly half the matches with combo rated teams, destabilizing the algo.

Is it really true that the "gap" is applied to individuals in doubles matches, rather than team?

My impression was that something like a 7.0/4.0 team vs. a 5.0/5.0 team would be treated like a 5.5 vs. 5.0 matchup in singles (average of each doubles team) and therefore weighted like a 0.5 gap matchup for everyone.

You seem to think it would be treated like a 2.0-gap for the 7.0 player and a 1.0-gap match for 4.0 player. Do you have evidence for that?

 

[travlerajm]

Is it really true that the "gap" is applied to individuals in doubles matches, rather than team?

My impression was that something like a 7.0/4.0 team vs. a 5.0/5.0 team would be treated like a 5.5 vs. 5.0 matchup in singles (average of each doubles team) and therefore weighted like a 0.5 gap matchup for everyone.

You seem to think it would be treated like a 2.0-gap for the 7.0 player and a 1.0-gap match for 4.0 player. Do you have evidence for that?

Let’s say it’s 8.0 mixed, with a 4.5M/3.5f vs 4.0/4.0 team.

The UTR combos for a typical match might be 8.5/3.0 vs. 5.0/4.0.

For the UTR 8.5 his opponent rating for this match would be 5 + 4 - 3 = 6.0. This means the gap is 8.5 - 6.0 = 2.5. So this match doesn’t count in the UTR algo, even though the ntrp combos for the two teams are exactly equal.

 

[TennisOTM]

Let’s say it’s 8.0 mixed, with a 4.5M/3.5f vs 4.0/4.0 team.

The UTR combos for a typical match might be 8.5/3.0 vs. 5.0/4.0.

For the UTR 8.5 his opponent rating for this match would be 5 + 4 - 3 = 6.0. This means the gap is 8.5 - 6.0 = 2.5. So this match doesn’t count in the UTR algo, even though the ntrp combos for the two teams are exactly equal.

But the high gap between the first two partners has nothing to do with that. The gap is 2.5 because you have an 11.5 team versus a 9.0 team.

If it was 5.75/5.75 vs. 5.0/4.0 it would come out the same way.

8.5 - (5 + 4 - 3) = 2.5
5.75 - (5 + 4 - 5.75) = 2.5

 

[jmnk]

Let’s say it’s 8.0 mixed, with a 4.5M/3.5f vs 4.0/4.0 team.

The UTR combos for a typical match might be 8.5/3.0 vs. 5.0/4.0.

For the UTR 8.5 his opponent rating for this match would be 5 + 4 - 3 = 6.0. This means the gap is 8.5 - 6.0 = 2.5. So this match doesn’t count in the UTR algo, even though the ntrp combos for the two teams are exactly equal.

I'd like to see a match where one partner is UTR8.5 and he is playing with UTR3.0. That would be awesome to see. Not from tennis perspective, from high comedy perspective. Besides, what would such match prove as far as one's level, no matter what the opposing pair is, and no matter what the score would be?

 

[Purestriker]

I'd like to see a match where one partner is UTR8.5 and he is playing with UTR3.0. That would be awesome to see. Not from tennis perspective, from high comedy perspective. Besides, what would such match prove as far as one's level, no matter what the opposing pair is, and no matter what the score would be?

That match happens more than you think in 8.0 mixed. From the ones I have watched, the 4.5 dominates and the 3.5W is very defensive. The 4.0's try to get every ball to the 3.5W, but it's not that easy against a 4.5M.

 

[schmke]

That match happens more than you think in 8.0 mixed. From the ones I have watched, the 4.5 dominates and the 3.5W is very defensive. The 4.0's try to get every ball to the 3.5W, but it's not that easy against a 4.5M.

And that is the "problem" (if you consider it a problem) in Mixed as the USTA presently runs it. You can have players a full 1.0 NTRP apart and then add in the generally accepted 0.5 gender difference making that effectively 1.5. Depending on how you slice it, that is a 4-5 UTR difference.

And just using UTRs conversion chart, they list a 3.5 female as 2.0-5.0 and a 4.5 male as 6.0-9.0, so the difference could be expected to be anywhere from 1.0 to 7.0.

 

[jmnk]

That match happens more than you think in 8.0 mixed. From the ones I have watched, the 4.5 dominates and the 3.5W is very defensive. The 4.0's try to get every ball to the 3.5W, but it's not that easy against a 4.5M.

How does 4.5M player look at himself in the mirror after such match (and we are assuming he is trying to win, and not play just for hits and giggles)??

 

[FuzzyYellowBalls]

How does 4.5M player look at himself in the mirror after such match (and we are assuming he is trying to win, and not play just for hits and giggles)??

I bet someone posting here can answer that question for you

 

[travlerajm]

Update:

I have hardly touched a racquet this month. Foul weather, lack of indoor courts, and stressful work situations taking away all my time to exercise.

But a prior opponent reached out to play this weekend. He was the age 30ish ex D3 college player solid high 4.5 ntrp and 7.xx 100% UTR.

Agreed to play him Saturday afternoon 3pm. Nice cool cloudy February day at forested Shoreview Park on the hillside, with only the clatter of the ever present pickleball crew to spoil the peace.

I knew my body wasn’t ready, but I needed the workout. In our previous match, we hadn’t finished within the 1:15 indoor time slot. He won first set 6-3, turn I was leading 3-2 up a break when we ran out of time.

Yesterday, I started fast to a 2-break 3-0 lead. Then he removed his fleece and and pulled to 3-2. So then I removed my sweatshirt and replaced my widebrim with the backward cap, and widened my lead to 5-2. Then he raised his game. I had 2 set points on his serve but couldn’t convert.

I had been serving only slice serves you that point yo save my unconditioned legs. But serving to close out the set at 5-4, and feeling my legs fading, I decided I needed a few “real” serves, and went hard and over the top. I was able use my serve to start the points at sn advantage. At 30-30, I pulled the trigger on a flat 2hb DTL, and the tennis gods smiled on my boldness and gave me a net cord winner. Then on 40-30 point I hit a well-executed fh approach to his bh corner, and got the crosscourt pass reply I expected. My bh volley stab putaway missed my target of the side T, barely clearing the net at a sharp angle 6” inside the deuce sideline. First set 6-4. high quality set.

Unfortunately my legs gave me the ‘no mas’ signal after that. I started serving underspin sidewinder drop serves, and the second set went fast 0-6.

I tried to see if my calves would allow me to spring into some “real” serves in the super, but my mobility was limited, as I felt like I was on the verge of busting an Achilles or a hammy. I went down meekly 10-4 in the super.

At least I got some good exercise. Sore today.

 

[Moon Shooter]

If a UTR 12 is really a UTR 12 and not a soft poser, he shouldn’t have much trouble with the challenge of double-bageling a UTR 8.

Even if we go along with assumption that a UTR 8 can take a game off a UTR 12, and it might even be expected, then it seems like the algo could easily find a way to handle it (by adjusting the expected score) rather than excluding the match altogether.

Yeah I don't know. Some players have a serve that, when it is clicking, will win games against pretty much anyone. The problem is it may not click as reliably as they like - and even lead to double faults - or the rest of their game may suffer because they can't move well or they have a weakness that many players exploit. I think the 2 point (or some margin) rule makes some sense.

 

[Max G.]

Yeah I don't know. Some players have a serve that, when it is clicking, will win games against pretty much anyone. The problem is it may not click as reliably as they like - and even lead to double faults - or the rest of their game may suffer because they can't move well or they have a weakness that many players exploit. I think the 2 point (or some margin) rule makes some sense.

I'm not entirely convinced of this. I think it's easy to underestimate how good players can be. I suspect that if, say, a 4.0 faces someone, thinks "man, this serve is so good, it could take points off of anyone!" but otherwise is competitive, then in reality a 5.0 could probably handily return that "4.0 but could take points off of anyone" serve even if it's clicking at 100%.

In my experience, at 4.5, I've never faced a serve that's so good that I think a 5.5 player would have trouble with it. I've faced plenty of serves that I can't deal with as a fellow 4.5, and some serves that would win games against 5.0 players, but having seen some good college players hit, who are probably 5.5... I don't think any of them would be troubled by even the best serves I've seen at 4.5.

 

[Moon Shooter]

I don’t have much experience at that level. But I have seen plenty of players in the 3.0 to low 3.5 hit unreturned serves against upper 4.0 players. You can see it in mixed doubles. Even at the pro level there are guys who can win games based on their serve. If anything that seems to be more pronounced at higher levels. One 3.5 player had a 120+ mph serves according to university of illinois radar system where we were playing some of our matches. His name was kevin vanmeter. That guy was new to tennis but I think used to be a pretty good pitcher or something like that. He was the most extreme example I saw but still I know some other, usually, tall players that it is night and day playing them when they have their serve going or not.

 

[ChaelAZ]

I don’t have much experience at that level. But I have seen plenty of players in the 3.0 to low 3.5 hit unreturned serves against upper 4.0 players. You can see it in mixed doubles. Even at the pro level there are guys who can win games based on their serve. If anything that seems to be more pronounced at higher levels. One 3.5 player had a 120+ mph serves according to university of illinois radar system where we were playing some of our matches. His name was kevin vanmeter. That guy was new to tennis but I think used to be a pretty good pitcher or something like that. He was the most extreme example I saw but still I know some other, usually, tall players that it is night and day playing them when they have their serve going or not.

I mean, if a 3.0 or 3.5 is able to hit enough unreturned serves against a 4.0, they will get enough games to be considered competitive at that level and move up as a 4.0. If they remain 3.0 or 3.5, they aren't able to serve consistently enough, or the 4.0 is able to handle it enough to get to the weaker ground game, and they can tell their friends they aced a higher level player x times in legendary 3.0 and 3.5 glory.

I think an example of what you are talking about of a lower rated player with a big serve is Trey at the Tennis Winners or ?? YT channel, who competes well enough against much higher rated players. I think he is only 3.5, but he'll be rated higher really quick. He also has developed much better groundies in the last year. It just works itself out really. He also has good athleticism. There aren't too many examples of *just* a big serve keeping a beginner rec player in matches against intermediate and advanced rec players.

It gets a little more cloudy in UTR 8+ range because no one at that level just relies on serving.

 

[Sea70]

When I signed up, I wrote in the “info about your level” field that I was a 4.5c, because the UTR ratings seem pretty useless for rec players who don’t play rated matches frequently.

When I signed up a few a weeks ago, my singles UTR was a 5. Today it has rebounded to 7.4. that you’ve

I just signed up to play in a UTR singles flex league for the first time.

As of today, I’m a UTR 5 in singles and a UTR 4 in doubles. Will I be able to make progress?

I will be reporting my opponent ratings and scores here.

Now that hove played for a while in utr recorded matches, how accurate is your utr?
What percentage of people are you winning with a lower utr than you? What percentage of people are you losing to with an utr higher than yours?

 

[Moon Shooter]

I mean, if a 3.0 or 3.5 is able to hit enough unreturned serves against a 4.0, they will get enough games to be considered competitive at that level and move up as a 4.0. If they remain 3.0 or 3.5, they aren't able to serve consistently enough, or the 4.0 is able to handle it enough to get to the weaker ground game, and they can tell their friends they aced a higher level player x times in legendary 3.0 and 3.5 glory.

I think an example of what you are talking about of a lower rated player with a big serve is Trey at the Tennis Winners or ?? YT channel, who competes well enough against much higher rated players. I think he is only 3.5, but he'll be rated higher really quick. He also has developed much better groundies in the last year. It just works itself out really. He also has good athleticism. There aren't too many examples of *just* a big serve keeping a beginner rec player in matches against intermediate and advanced rec players.

It gets a little more cloudy in UTR 8+ range because no one at that level just relies on serving.

I haven’t seen the people your talking about. But yes people with big serves but are not consistent. I play against a guy who it seems 3/4 of his serves are either unreturned or double faults. If fewer double faults it means he has a good day, more means a bad day and in between means typical. Also I think in my area we tend not to only play people at our level in mixed usta and in social doubles. I do not have a big serve but even I as a 3.0 can chalk up a game against a 4.0 every now and then.


The question isn’t whether they are “keping in matches.” The question is whether they can win a game or two in the match and drag down the rating of the player over 2 utr higher.

As far as people over 8 utr relying on their serve consider these percentages of service games won for top pro servers. F anything the advantage of the serve seems more pronounced as people improve.

 

[FuzzyYellowBalls]

Now that hove played for a while in utr recorded matches, how accurate is your utr?
What percentage of people are you winning with a lower utr than you? What percentage of people are you losing to with an utr higher than yours?

100%, 100%, and 100% is the answer.

Problem is it's a little UTR bubble with people playing only flex league and churning each other's ratings (think small round robin). The ratings are 100% though, so mission accomplished for TJ and it's not his fault it's a small bubble. Imagine though if you grabbed 8 people or so and just had them play each other over and over, a UTR would result eventually. I wonder though if you are happy with your rating TJ or not so much?

 

[schmke]

100%, 100%, and 100% is the answer.

Problem is it's a little UTR bubble with people playing only flex league and churning each other's ratings (think small round robin). The ratings are 100% though, so mission accomplished for TJ and it's not his fault it's a small bubble. Imagine though if you grabbed 8 people or so and just had them play each other over and over, a UTR would result eventually. I wonder though if you are happy with your rating TJ or not so much?

Perhaps UTR (or any other rating system that shows a confidence) should factor in how connected the player or their opponents are with the rest of the ecosystem and not display 100% until some "degrees of separation" threshold is met?

 

[FuzzyYellowBalls]

Perhaps UTR (or any other rating system that shows a confidence) should factor in how connected the player or their opponents are with the rest of the ecosystem and not display 100% until some "degrees of separation" threshold is met?

Sounds good to me, but also a huge headache .

 

[travlerajm]

Now that hove played for a while in utr recorded matches, how accurate is your utr?
What percentage of people are you winning with a lower utr than you? What percentage of people are you losing to with an utr higher than yours?

100%, 100%, and 100% is the answer.

Problem is it's a little UTR bubble with people playing only flex league and churning each other's ratings (think small round robin). The ratings are 100% though, so mission accomplished for TJ and it's not his fault it's a small bubble. Imagine though if you grabbed 8 people or so and just had them play each other over and over, a UTR would result eventually. I wonder though if you are happy with your rating TJ or not so much?

I’ve been busier and playing fewer matches and watching my UTR rating decay between matches, just like old times.

 

[TennisOTM]

Problem is it's a little UTR bubble with people playing only flex league and churning each other's ratings (think small round robin).

Not true though, only a couple of his opponents are UTR-only players. Most of them have USTA singles tournament and/or league matches on their record, a few of them extensive. I'm guessing his singles rating is pretty solid right now, although a couple more matches against established players near or above his current rating would help.

 

[travlerajm]

Not true though, only a couple of his opponents are UTR-only players. Most of them have USTA singles tournament and/or league matches on their record, a few of them extensive. I'm guessing his singles rating is pretty solid right now, although a couple more matches against established players near or above his current rating would help.

Except that I’m currently only touching a racquet once per week due to increased commitments outside of rec tennis, so playing more solid players is unliking to improve my rating.

 

[FuzzyYellowBalls]

Not true though, only a couple of his opponents are UTR-only players. Most of them have USTA singles tournament and/or league matches on their record, a few of them extensive. I'm guessing his singles rating is pretty solid right now, although a couple more matches against established players near or above his current rating would help.

Except that I’m currently only touching a racquet once per week due to increased commitments outside of rec tennis, so playing more solid players is unliking to improve my rating.

At least one thing is constant, UTR remains @travlerajm least favorite rating system? All is right in the universe.

 

[TennisOTM]

Except that I’m currently only touching a racquet once per week due to increased commitments outside of rec tennis, so playing more solid players is unliking to improve my rating.

By "help" I meant help improve the rating's accuracy (predictive power), which of course could send it in either direction.

If someone's "true" ability swings dramatically from month-to-month because of irregular practice, that's going to be tough for any rating system to account for.

 

[travlerajm]

By "help" I meant help improve the rating's accuracy (predictive power), which of course could send it in either direction.

If someone's "true" ability swings dramatically from month-to-month because of irregular practice, that's going to be tough for any rating system to account for.

In that sense, my UTR rating’s accuracy is improving.

 

[Moon Shooter]

Perhaps UTR (or any other rating system that shows a confidence) should factor in how connected the player or their opponents are with the rest of the ecosystem and not display 100% until some "degrees of separation" threshold is met?

I suspect the vast majority of UTR's adult rec matches come from USTA competition. In my area we have no UTR events for adult rec players at all so it's all USTA events. That means they will always have the separate ecosystems that USTA creates. Namely separate systems for men and women and adult and school related play. (possibly regional as well)

It does seem that UTR is connecting men and women's ratings much better then the World Tennis Number USTA is promoting. I'm convinced that USTA is manually manipulating the WTN ratings to make men's ratings worse then women's ratings for the same skill level. UTR seems the most accurate published rating system if you have at least 10 matches in the category (singles or doubles) in the past 12 months.

 

[TennisOTM]

I just noticed that UTR singles flex leagues are now offering prize money for UTR 7+ participants. This could add some serious spice to this blog - will you go for it??

From the league description:

UTR 7+ Prize money
UTR 7+ only. If you are unrated, the decision is 100% on UTR sports to decide if you are eligible or not. Players will be split into groups of 10 players (when possible). If two players end up with the same record, it will come down to head to head followed by set one, followed by games won to decide who is the winner. The prize money breakdown is below (winner take all)
5-10 players: $150
10-12 players: $200
12-19 players: $250, split between two groups ($125 for each group winner)
20: $400 split between two groups ($200 for each group winner)

 

[Moon Shooter]

But the high gap between the first two partners has nothing to do with that. The gap is 2.5 because you have an 11.5 team versus a 9.0 team.

If it was 5.75/5.75 vs. 5.0/4.0 it would come out the same way.

8.5 - (5 + 4 - 3) = 2.5
5.75 - (5 + 4 - 5.75) = 2.5

His example shows why mixed doubles teams below 9.0 tend to get more skill on the court by playing a higher USTA level male with a lower USTA level female.

 

[tennis3]

I just noticed that UTR singles flex leagues are now offering prize money for UTR 7+ participants. This could add some serious spice to this blog - will you go for it??

From the league description:

UTR 7+ Prize money
UTR 7+ only. If you are unrated, the decision is 100% on UTR sports to decide if you are eligible or not. Players will be split into groups of 10 players (when possible). If two players end up with the same record, it will come down to head to head followed by set one, followed by games won to decide who is the winner. The prize money breakdown is below (winner take all)
5-10 players: $150
10-12 players: $200
12-19 players: $250, split between two groups ($125 for each group winner)
20: $400 split between two groups ($200 for each group winner)

How much is the registration fee? I assume it had to go up by $15 or so to cover the prize money?

 

[travlerajm]

This blog has been inactive for awhile, mostly because I haven’t played any tennis for the last couple months, with stressful other stuff going on.

But I motivated myself to register for the next UTR singles flex league session that starts in a week.

Win or lose, the stories will resume.

 

[travlerajm]

Update:
A couple months ago, I joined a usta 4.5 singles flex league for the first time.

One guy reached out to me, but the match never happened due to weather and scheduling issues. And since I’ve been away from the tennis court, I haven’t been pro active. So I’ve played zero league matches.

Last night, another random dude from the usta 4.5 singles flex league popped up and wants to schedule a match. We’ve coordinated and scheduled our match for outdoor Thursday night.

Since it’s been like 6 weeks since I’ve played any sort of singles, my legs are in no condition to last a full singles match. I got back into the racquetball court this week to build a new type of racquet to prep for singles.

I’ll need to bring out all the technology stops to have a chance.

 

[StringSnapper]

This blog has been inactive for awhile, mostly because I haven’t played any tennis for the last couple months, with stressful other stuff going on.

But I motivated myself to register for the next UTR singles flex league session that starts in a week.

Win or lose, the stories will resume.

Hope everything is OK. For as much **** as I've given you for your writing style, it provides a lot of entertainment and hope you get back on the court soon. I've been injured for over 6 months now with a knee surgery planned so need to live vicariously through someone!

 

[TennisOTM]

How much is the registration fee? I assume it had to go up by $15 or so to cover the prize money?

The registration fee is $30 for subscibers and $42 for non-subscribers. That's the same fee as the prior month, when there was no prize money offered. I think the prize money for a few UTR 7+ groups will be more than covered by the fees from the lower-level groups.