NTRP vs. UTR Sandbagger Ratings

[BallBag]

So I was trolling through the ratings of some known 4.0 sandbaggers in my area and noticed that although they are managing their Dynamic NTRP (under 3.95 on TR) their UTRs are closer 7.0. Since both rating are feeding from the same results pool, how is UTR closer to the actual level?
I have no knowledge of either rating algorithm (paging Dr. @schmke ) so I can only speculate that UTR's weighing down mismatched opponents (diff greater than 2 UTR) causes the sandbaged results to have less effect.
I'm basing this on evidence from a sample size of two and my personal intuition. Has anyone else noticed UTR better at flagging sandbaggers?

 

[jm1980]

TennisRecord is not accurate, and UTR is pretty shaky at this level

I wouldn’t trust either

 

[BallBag]

TennisRecord is not accurate, and UTR is pretty shaky at this level

I wouldn’t trust either

UTR is pretty good when applied within a group and TR was right about most of last year's bump ups (even before they applied the correction). This is just my observation for the small group players I tracked.

 

[Ft.S]

TR is so way off, it's ratings are useless from the NTRP perspective. UTR is tracking the players in my general area far more accurately and is a better indicator of playing performance than TR. IMHO, the ultimate NTRP estimator is @schmke, but his last numbers for me are beyond surprising, although he has been proven right far more often than not, we'll see

 

[J011yroger]

UTR has had me from barely 9 to almost 11 this year, it goes up or down over a full point without me playing a single match.

J

 

[BallBag]

UTR has had me from barely 9 to almost 11 this year, it goes up or down over a full point without me playing a single match.

J

Is it all USTA results?

 

[J011yroger]

Is it all USTA results?

Yea, leagues and tournaments.

I've played both, juniors and college players are rated higher than league players and women are rated higher than men, so if I only played USTA league I would top out around 10, if I only played Men's Open I'd be an 11, and if I could play in women's college I'd be a 12 lol.

J

 

[ShaunS]

You could argue that it's relatively easy to identify someone who is putting up terrible results to get bumped down, but those guys will be easy to spot when they win matches.

The real difficulty comes when you're hoping to identify the guys who are "shaving points". It's going to be incredibly hard for any system to spot when this is happening because of how variable humans are in regular situations. I don't think any rating system is going to be able to accurately work around those "bad" results in the near future. With enough data you could hope to derive some level of competency, but there's no telling where that is today.

And for what it's worth, I would agree that TR's accuracy is not nearly so good that I'd think a 3.95 isn't at risk. I just looked at my summer 4.0 team. There's a guy who's 3.97 dynamic, and I'll be dumbfounded if he doesn't get bumped to 4.5 at the end of year. For that matter, I'm expecting someone who's 3.94 to get bumped too.

 

[schmke]

So I was trolling through the ratings of some known 4.0 sandbaggers in my area and noticed that although they are managing their Dynamic NTRP (under 3.95 on TR) their UTRs are closer 7.0. Since both rating are feeding from the same results pool, how is UTR closer to the actual level?
I have no knowledge of either rating algorithm (paging Dr. @schmke ) so I can only speculate that UTR's weighing down mismatched opponents (diff greater than 2 UTR) causes the sandbaged results to have less effect.
I'm basing this on evidence from a sample size of two and my personal intuition. Has anyone else noticed UTR better at flagging sandbaggers?

Why do you assume UTR is closer to the actual level? They are different algorithms with different goals and will produce different output potentially from the same match results.

But, UTR itself says their range for an NTRP 4.0 is UTR 4.5-7.0, and an NTRP 4.5 is 6.0-8.5 (each a full whopping 6 levels), so even if you assume TR is correct (which as others have pointed out is questionable), a borderline NTRP 4.0/4.5 being a UTR 7.0 is more or less what you'd expect.

 

[J011yroger]

Why do you assume UTR is closer to the actual level? They are different algorithms with different goals and will produce different output potentially from the same match results.

But, UTR itself says their range for an NTRP 4.0 is UTR 4.5-7.0, and an NTRP 4.5 is 6.0-8.5 (each a full whopping 6 levels), so even if you assume TR is correct (which as others have pointed out is questionable), a borderline NTRP 4.0/4.5 being a UTR 7.0 is more or less what you'd expect.

Am I the only one who wonders how they made that conversion chart since it doesn't line up with reality?

J

 

[schmke]

Am I the only one who wonders how they made that conversion chart since it doesn't line up with reality?

J

Make the UTR range for each NTRP level really big and overlapping to A) ensure your conversion chart is never wrong, and B) set the perception that NTRP is wildly inaccurate if one level spans such a large range of an "accurate" algorithm?

Of course, if you correspondingly create a conversion chart going from UTR to NTRP, you'd also have to end up having a large range to never be wrong and would set the perception that UTR is wildly inaccurate if one of their levels spans such a large range of NTRP levels ...

 

[J011yroger]

Make the UTR range for each NTRP level really big and overlapping to A) ensure your conversion chart is never wrong, and B) set the perception that NTRP is wildly inaccurate if one level spans such a large range of an "accurate" algorithm?

Of course, if you correspondingly create a conversion chart going from UTR to NTRP, you'd also have to end up having a large range to never be wrong and would set the perception that UTR is wildly inaccurate if one of their levels spans such a large range of NTRP levels ...

All I'm saying is that no good 4.5 is below 8.5utr and no 5.5 is anywhere near 10 or 11.

J

 

[BallBag]

Why do you assume UTR is closer to the actual level? They are different algorithms with different goals and will produce different output potentially from the same match results.

But, UTR itself says their range for an NTRP 4.0 is UTR 4.5-7.0, and an NTRP 4.5 is 6.0-8.5 (each a full whopping 6 levels), so even if you assume TR is correct (which as others have pointed out is questionable), a borderline NTRP 4.0/4.5 being a UTR 7.0 is more or less what you'd expect.

I was looking at guys I know who are legit upper third of 4.0 and they are all low 6. The two sandbaggers were at low 7.

 

[J011yroger]

All I'm saying is that no good 4.5 is below 8.5utr and no 5.5 is anywhere near 10 or 11.

J

And that's with playing USTA that hurts your rating.

J

 

[BallBag]

Am I the only one who wonders how they made that conversion chart since it doesn't line up with reality?

J

UTR has access to a data base of everyone's UTR and NTRP rating. I'm guessing they did a normal distribution of UTRs for every NTRP bracket.

 

[J011yroger]

UTR has access to a data base of everyone's UTR and NTRP rating. I'm guessing they did a normal distribution of UTRs for every NTRP bracket.

J

 

[BallBag]

J

Why? If they can get your scores from TennisLink then they can also get your NTRP level. It looks like they made a histogram with 0.25 bins for every NTRP level. There's a wide range in the levels because there is an acctual wide range in abilities. There's an overlap because there is an actual overlap.

 

[J011yroger]

Why? If they can get your scores from TennisLink then they can also get your NTRP level. It looks like they made a histogram with 0.25 bins for every NTRP level. There's a wide range in the levels because there is an acctual wide range in abilities. There's an overlap because there is an actual overlap.

Because their chart doesn't line up with reality.

J

 

[MathGeek]

Why? If they can get your scores from TennisLink then they can also get your NTRP level. It looks like they made a histogram with 0.25 bins for every NTRP level. There's a wide range in the levels because there is an acctual wide range in abilities. There's an overlap because there is an actual overlap.

They can only get NTRP levels for adults. Junior NTRP levels are unavailable to the general public.

 

[J011yroger]

They can only get NTRP levels for adults. Junior NTRP levels are unavailable to the general public.

Isn't that what we are talking about?

J

 

[BallBag]

Because their chart doesn't line up with reality.

J

I'm gonna look into that.

 

[MathGeek]

Isn't that what we are talking about?

J

My point is that when making a UTR to NTRP comparison chart, a lot of the data (Junior NTRPs) is missing.

 

[BallBag]

They can only get NTRP levels for adults. Junior NTRP levels are unavailable to the general public.

They didn't post conversions for Junior NTRP to UTR.

Edit: just saw your reply. They did a conversion for sectional and national junior tournament participants.

 

[MathGeek]

They didn't post conversions for Junior NTRP to UTR.

Did they specify "Adult NTRP" or did you just assume "Adult"?

 

[BallBag]

Did they specify "Adult NTRP" or did you just assume "Adult"?

Assumed adult but I'm pretty sure that the whole Junior NTRP thing has been defunct. Correct me if I'm wrong.

 

[heninfan99]

UTR has had me from barely 9 to almost 11 this year, it goes up or down over a full point without me playing a single match.

J

Can a UTR expert explain this...???

 

[J011yroger]

Did they specify "Adult NTRP" or did you just assume "Adult"?

Yes, that's why it says Men's as opposed to boy's.

J

 

[BallBag]

Can a UTR expert explain this...???

Jolly beat a bunch of 12 UTR rated teenagers about a year ago. Since then he swore off child abuse and been playing USTA league where he is about even playing against 10 UTR rated adults. UTR sunset the child abuse results which ignored those wins and lowered his UTR.

(Disclaimer: This is wild theorization based entirely on my imagination. Also, same thing happened to be but with half a UTR point. )

 

[CosmosMpower]

A UTR 7 is not a 4.0 sandbagger it's a very high 3.5 sandbagger. UTR 7 is a mid/high 4.0. A sandbagger 4.0 is like a UTR 9 or 10.

 

[CosmosMpower]

Can a UTR expert explain this...???

A whole point is a lot but it can go up and down based on your previous opponents activity and current rating.

 

[Deleted member 23235]

So I was trolling through the ratings of some known 4.0 sandbaggers in my area and noticed that although they are managing their Dynamic NTRP (under 3.95 on TR) their UTRs are closer 7.0. Since both rating are feeding from the same results pool, how is UTR closer to the actual level?
I have no knowledge of either rating algorithm (paging Dr. @schmke ) so I can only speculate that UTR's weighing down mismatched opponents (diff greater than 2 UTR) causes the sandbaged results to have less effect.
I'm basing this on evidence from a sample size of two and my personal intuition. Has anyone else noticed UTR better at flagging sandbaggers?

i’d guess it’s because UTR ignores matches where the rating difference is really big


Sent from my iPhone using Tapatalk

 

[BallBag]

A UTR 7 is not a 4.0 sandbagger it's a very high 3.5 sandbagger. UTR 7 is a mid/high 4.0. A sandbagger 4.0 is like a UTR 9 or 10.

Ya'll have some high speed 4.0s. The 4.0s that went to Nationals were around 7 UTR.

 

[Vox Rationis]

You'll have some high speed 4.0s. The 4.0s that went to Nationals were around 7 UTR.

Excuse him. He's from Texas so his perception's a little skewed. /s

A UTR 7 is not a 4.0 sandbagger it's a very high 3.5 sandbagger. UTR 7 is a mid/high 4.0. A sandbagger 4.0 is like a UTR 9 or 10.

I skimmed through the UTRs of players at nationals when this was brought up in another thread. The 4.0s at nationals were typically UTR 6 with the best ones being a UTR 7. There were a couple at a UTR 8 but they all had estimated NTRPs well over 4.0. I dare you to find a 4.0 with a UTR of 10. That's 5.0 range with only elite 4.5s (shout out to Jolly) hitting that mark.

At 3.5 nationals, the good 3.5s were only at a UTR 5, including players on the semifinalist teams. I checked the sandbaggers who were winning 0 & 0 and 1 & 1 most of the time and they were almost all at a UTR 6. There were only two that I saw, out of all the best 3.5 national players that were a UTR 7. One of them was a guy on Midwests team who not only dominated 3.5 all year but also went undefeated at 4.0 (4-0 record). I would not be surprised to see him double bumped to 4.5.

 

[Deleted member 293577]

UTR is not very accurate in my experience. TLS and TR are closer to reality.

 

[JLyon]

UTR for a 4.5 solid player probably should be 9-10, D1 Legit College Players for Guys is 11.5+
Ladies D1 College 9.0+
4..0 probably 6.5-8.5 range.
I am a low 4.5 now days and local college girls are UTR 9.0-10 level and I have issues hanging even close with them now days in singles

 

[CosmosMpower]

Excuse him. He's from Texas so his perception's a little skewed. /s


I skimmed through the UTRs of players at nationals when this was brought up in another thread. The 4.0s at nationals were typically UTR 6 with the best ones being a UTR 7. There were a couple at a UTR 8 but they all had estimated NTRPs well over 4.0. I dare you to find a 4.0 with a UTR of 10. That's 5.0 range with only elite 4.5s (shout out to Jolly) hitting that mark.

At 3.5 nationals, the good 3.5s were only at a UTR 5, including players on the semifinalist teams. I checked the sandbaggers who were winning 0 & 0 and 1 & 1 most of the time and they were almost all at a UTR 6. There were only two, out of all the best 3.5 national players that were a UTR 7. One of them was a guy on Midwests team who not only dominated 3.5 all year but also went undefeated at 4.0 (4-0 record). I would not be surprised to see him double bumped to 4.5.

I am a UTR 6, 3.5 and the team that beat us in the semifinals of TX sectionals had two UTR 7 singles players and the rest of the top 8 were pretty much all UTR 6's (one UTR 7 dubs). They didn't even get to nationals, lost to the Houston team. I'm surprised to hear nationals 3.5 teams only had UTR 5's.

 

[BallBag]

I am a UTR 6, 3.5 and the team that beat us in the semifinals of TX sectionals had two UTR 7 singles players and the rest of the top 8 were pretty much all UTR 6's (one UTR 7 dubs). They didn't even get to nationals, lost to the Houston team. I'm surprised to hear nationals 3.5 teams only had UTR 5's.

Ya'll have some high speed sandbaggers.

 

[Vox Rationis]

I am a UTR 6, 3.5 and the team that beat us in the semifinals of TX sectionals had two UTR 7 singles players and the rest of the top 8 were pretty much all UTR 6's (one UTR 7 dubs). They didn't even get to nationals, lost to the Houston team. I'm surprised to hear nationals 3.5 teams only had UTR 5's.

Not sure if this is allowed but it's all public info and only took me 5 minutes to look up copy and paste so here's the finalist teams UTRs listed by which rating was higher/what they tended to play more of

Intermountain
Singles:
A P - 6
L L C S - 6

Doubles:
T C - 7
D E - 7
B B - 4
J V - 5
Z S - 6
B G - 6

Nor Cal
Singles:
J P - 4
V H Q N - 5

Doubles:
T B - 6
H T - 6
q n - 5
N D P - 6
T X P - 5
D N - 5

I think teams like Intermountain and the ones you saw in texas are more the exception than the rule.

 

[BallBag]

Not sure if this is allowed but it's all public info and only took me 5 minutes to look up copy and paste so here's the finalist teams UTRs listed by which rating was higher/what they tended to play more of

Intermountain
Singles:
A P - 6
L L C S - 6

Doubles:
T C - 7
D E - 7
B B - 4
J V - 5
Z S - 6
B G - 6

Nor Cal
Singles:
J P - 4
V H Q N - 5

Doubles:
T B - 6
H T - 6
q n - 5
N D P - 6
T X P - 5
D N - 5

I think teams like Intermountain and the ones you saw in texas are more the exception than the rule.

Those two 7s bag some serious sand. They went 60 62 and 61 64 against UTR 6s but also 64 64 and 64 63 against UTR 3s. Or maybe they took some lessons during the season.

 

[J011yroger]

Don't post people's real names, that's just dumb.

J

 

[CosmosMpower]

Don't post people's real names, that's just dumb.

J

Agree but it is public info....

 

[CosmosMpower]

Ya'll have some high speed sandbaggers.

True, I just played a UTR 9 player in 3.5 fall doubles yesterday. 3.5 self rated lol.

 

[ShaunS]

Agree but it is public info....

Technically, a lot of stuff is publicly available information, such as your home address. Doesn't mean it's right to post it.

 

[Vox Rationis]

Don't post people's real names, that's just dumb.

J

Technically, a lot of stuff is publicly available information, such as your home address. Doesn't mean it's right to post it.

I agree, which is why I've edited it. I do think it's a bit different when it's just a name grabbed from a public source without any other info attached to it. But regardless, names don't belong on here. Stupid choice on my part.