Ratings Madness......

[michael_1265]

This player played for me when I captained the 3.0 men’s team until 2010. With the 2011 season 3.0 team in question, he joined a 3.5 team. I am providing his 2010 record to show that his 3.0 dynamic rating was likely low at the end of 2010.

As a 3.0 (90% doubles)
2010 3.0 Men’s 2-7, with a total of 5 sets won
2010 Senior 3.5 Men’s 2-6, with a total of 4 sets won
2011 3.5 Men’s: 0-11, with no sets won. Most of the matches were not competitive.

He was bumped to 3.5 in November.

This is the most convincing proof I’ve seen that the algorithm is biased toward merely playing at a higher level, but not necessarily being competitive.

Oddly, I know a very skilled 4.0 who had a bad season last year and was bumped down to 3.5, but his record was not nearly as bad as 0-11.

Once more, my theory that the USTA uses a giant Magic 8-ball to make decisions gains credence.

 

[gameboy]

Not very convincing to me.

First, he did have 2 wins in senior 3.5 Men's. That may be enough to rate him at 3.5

Second, as countless people around here have said it over and over again, W/L does not matter. What matters is how many games the person won during those matches and the ratings of the person playing against him.

3-3 loss against a benchmarked 3.5 is going to register as a competitive 3.5 player. It does not matter if he plays the same guy 10 times and lose 10 times with similar margins.

If he is not competitive at all in 3.5, he will be bumped down.

 

[SweetH2O]

Meanwhile, I have a 3.5 self rating at end of 2010, get a 4.0 early-start rating halfway through 2011, and play in a 4.0 league through the rest of the year. I have an over .500 record in 4.0 league, but remain 3.5 at year end 2011.

My hypothesis is that when they try to standardize the levels nationwide, some sections will be hard to be bumped up if your section was seen as "weak", and other sections will be easier if they are seen as "strong". But I'm just guessing like everyone else.

 

[J_R_B]

Meanwhile, I have a 3.5 self rating at end of 2010, get a 4.0 early-start rating halfway through 2011, and play in a 4.0 league through the rest of the year. I have an over .500 record in 4.0 league, but remain 3.5 at year end 2011.

My hypothesis is that when they try to standardize the levels nationwide, some sections will be hard to be bumped up if your section was seen as "weak", and other sections will be easier if they are seen as "strong". But I'm just guessing like everyone else.

This is almost certainly true with you. The early start ratings are run before the "nationals correction", the year-end after. Therefore, your results alone warranted a bump, but apparently, your sectional correction un-bumped you.

 

[OrangePower]

2011 3.5 Men’s: 0-11, with no sets won. Most of the matches were not competitive.

You'd need to clarify what you view as 'not competitive'. For example, playing against a top level 3.5, a score of 2 and 2 would be considered competitive enough to warrant a low 3.5 rating. Remember that according to USTA, a person at top of level should theoretically bagel/breadstick a person at bottom of level.

 

[michael_1265]

You'd need to clarify what you view as 'not competitive'. For example, playing against a top level 3.5, a score of 2 and 2 would be considered competitive enough to warrant a low 3.5 rating. Remember that according to USTA, a person at top of level should theoretically bagel/breadstick a person at bottom of level.

I always thought two breaks or more was considered non-competitive.

 

[OrangePower]

^^^^

I'd agree with that as a casual definition, but I think what the algorithm considers 'competitive' is more like 1 & 2.

Just to be clear, there is no specific 'competitiveness' calculation in the algorithm per se, but what the algorithm does is correlate between score and expected difference in DNTRP. My estimate (who the heck knows for sure) is that a score of 1 & 2 (maybe even 1 & 1) translates to an expected DNTRP difference of 0.5 (i.e. one full level). Just based on experience (observed records of people being bumped and so on), and the various statements USTA drops on us from time to time (e.g. what ranges of scores to reasonably expect within level).

In the case you mentioned, if the player in question was getting 2-3 games per match against *top* 3.5's, that would explain the bump up.

 

[michael_1265]

^^^^

I'd agree with that as a casual definition, but I think what the algorithm considers 'competitive' is more like 1 & 2.

Just to be clear, there is no specific 'competitiveness' calculation in the algorithm per se, but what the algorithm does is correlate between score and expected difference in DNTRP. My estimate (who the heck knows for sure) is that a score of 1 & 2 (maybe even 1 & 1) translates to an expected DNTRP difference of 0.5 (i.e. one full level). Just based on experience (observed records of people being bumped and so on), and the various statements USTA drops on us from time to time (e.g. what ranges of scores to reasonably expect within level).

In the case you mentioned, if the player in question was getting 2-3 games per match against *top* 3.5's, that would explain the bump up.

I see your point. From what I can see, most of the competition he has faced has been middle of the road, but a couple of the better teams might be in the mix. It's really difficult, though, to imagine that someone is going to be competitive when they have never even been .500 at the lower level. I am a middling 3.5. I play regularly with a former Div I player. 50% of the sets are 6-0, but I often squeak 1, and sometime 2 games in a set. I can often get to 3 or more games against good 4.0s. I never beat them, and I guess that's the point. Winning really is the acid test. Going 0-11 and getting bumped up is like a pro player being #1 in the world for two years without winning a major........and we know that never happens:twisted:

 

[equinox]

So appeal. Point out the 0-11 record.

 

[michael_1265]

So appeal. Point out the 0-11 record.

I think that he wants to be a 3.5. I am worried that it will be demoralizing, but I guess if he can deal with 0-11, I shouldn't worry.

The point of the thread was how the algorithm treats playing up.

 

[OrangePower]

I see your point. From what I can see, most of the competition he has faced has been middle of the road, but a couple of the better teams might be in the mix. It's really difficult, though, to imagine that someone is going to be competitive when they have never even been .500 at the lower level. I am a middling 3.5. I play regularly with a former Div I player. 50% of the sets are 6-0, but I often squeak 1, and sometime 2 games in a set. I can often get to 3 or more games against good 4.0s. I never beat them, and I guess that's the point. Winning really is the acid test. Going 0-11 and getting bumped up is like a pro player being #1 in the world for two years without winning a major........and we know that never happens:twisted:

Winning is the acid test, but too course grained to use as the only parameter in a rating algorithm. If you played against a large enough sample of players then it would work, but many players play just a limited number of matches each season that 'count' towards DNTRP.

Maybe some combination of the current method together with a sets won/lost type formula would work best, but nothing is going to be perfect.

At the end of the day, playing up greatly increases your chances of being bumped even if your results are mediocre, and really I have no problem with that. By playing up you are essentially representing yourself as being good enough to hang with those at the higher level, so there you go.

 

[Angle Queen]

The point of the thread was how the algorithm treats playing up.

I dunno, Michael. You and I have had this discussion before. Playing up, in and of itself, I don't think even enters into the algorithm. As Orange Power points out, though, you'll generally be playing higher rated players...and therefore, one would think even moderate success against would give your DNTRP a boost.

No, madness is some of the people on your neighbor's team that went all the way to the National finals...and didn't get bumped. You know they were already top of their class because that's why they were recruited for that SuperTeam in the first place.

For computer-rated players, I'd give up trying to figure out The Computer. Save your energies for ferreting out and complaining about inappropriate self-rates.

 

[gmatheis]

The problem with playing up is that a high 3.5 may feel bad beating someone 6-0 6-0 (I know I feel bad bageling opponents) and let them have a couple games each set. By doing that though the USTA now thinks the lower player "earned" those 2 games and gives him more credit than he deserves.

 

[Angle Queen]

The problem with playing up is that a high 3.5 may feel bad beating someone 6-0 6-0 (I know I feel bad bageling opponents) and let them have a couple games each set. By doing that though the USTA now thinks the lower player "earned" those 2 games and gives him more credit than he deserves.

Here's my lesson learned (I suppose) -- twice this past season when I played (singles) in-level, my opponents were at the lower end of 3.5 (either a new bump or admittedly on the decline)....and I took it easy, tried out some new things. On paper, neither one took a set's worth of games from me in total. Still, I wonder if had I kept my foot on the gas if my year-end NTRP would have been different.

There might be those who say you should always press for the biggest win since flight-winner might be determined by sets or games (ours actually was last year :shock: )...but there's no sense in making the experience miserable for everyone. In-level beatdowns will really screw with your mind and neither of those ladies deserved that.

 

[gameboy]

Playing up and complaining about a bump up is like complaining about getting herpes after binging on prostitutes.

If you don't want to be bumped up, don't play up...

 

[ian2]

I'd put it this way: if a player wants to be bumped up, it's much easier to do so by playing up than by playing at level. In fact, a bump-up is almost guaranteed when playing up, nearly irrespective of the actual results.

This is due to the way the algorithm works: a player's DNTRP will almost always go up when playing someone with a substantially higher DNTRP (by "substantially higher" I mean 0.25+ range) - even if the "lower" player lost to the tune of 6-2 6-3. While that same player beating someone at-level may not raise their DNTRP; it may in fact lower it. In other words, for a player at the higher range of a given rating, it's very difficult to raise their DNTR by playing at-level, as this basically requires blow-out wins against players at the lower range of that rating, and substantial (score-wise) wins over players with similar ratings.

My results are a case-in-point. 2010: 0.750 W/L at 3.5, with lots of singles wins in 6-2 6-1 range = no bump-up. 2011 - playing up exclusively: 0.350 W/L, with most wins in competitive matches, and a few less-than competitive (score-wise) losses = bump up.

Exact same picture for another guy on my team who followed the same route: excellent results in 2010 at 3.5, no bump-up. Middling results playing up at 4.0, bump-up.

Yet another guy on my team who was a mid-level 3.5 based on 2010 results. He also played up exclusively in 2011. Only played four of five matches, all doubles with a strong 4.0 partner. Lost all matches, some with competitive scores and some not. End result? Bumped up.

Bottom line: don't play at level if you want to be bumped-up; play up exclusively. And vice-versa: don't play up if you don't want to be bumped up; you might still be bumped up but you'd have to perform quite well for that to happen. A caveat: in the last 2-3 years USTA was making bulk "post-algorithm" adjustments to the year-end ratings. Specifics differed year-to-year as well as (it would seem) regionally, but overall the effect was of aggressively moving up players at the lower NTRP ranges. That was the case for 2.5, 3.0, and 3.5 in 2009, and for 2.5 and 3.0 in 2010 and 2011.

 

[backttennis]

Does anyone know if moving to a different part of the country in between early start and year end ratings could affect your rating? I just got my early start rating, and will be snowbirding, i.e. moving from New England to Florida in early November and changing my address immediately to sign up to play in USTA singles flex leagues in Florida.

Flex league play does not count at all towards NTRP ratings, but that will be all I'm interested in doing as far as USTA play.

 

[floridatennisdude]

Playing up and complaining about a bump up is like complaining about getting herpes after binging on prostitutes.

If you don't want to be bumped up, don't play up...

Analogy of the week nominee

 

[Angle Queen]

+1

Ya know, FTD, I missed that one.

But more to backttennis's bump, I don't have a clue. We don't have ESR here so hopefully some of the FLA gang can comment.

 

[J_R_B]

Does anyone know if moving to a different part of the country in between early start and year end ratings could affect your rating? I just got my early start rating, and will be snowbirding, i.e. moving from New England to Florida in early November and changing my address immediately to sign up to play in USTA singles flex leagues in Florida.

Flex league play does not count at all towards NTRP ratings, but that will be all I'm interested in doing as far as USTA play.

No. The adjustments due to nationals are based on who you played in the previous year, not where you currently live. Flex leagues don't count for NTRP.