Amarone or another ratings expert...???

[HoustonHacker]

I made it to Sectionals this year in 4.0 and am therefore a benchmark player. My friend wants me to play some tournaments with him at the 5.0 and 4.5 level. It would be fun, but I wouldn't want to risk being "bumped up" as a result.

My question: I've heard it said that as a benchmark my rating will be determined through league play and then I will be "used" to help calculate others' ratings if I should play tournaments. Is this true? Will my tournament results affect my rating at all or is it already set in stone?

 

[raiden031]

I made it to Sectionals this year in 4.0 and am therefore a benchmark player. My friend wants me to play some tournaments with him at the 5.0 and 4.5 level. It would be fun, but I wouldn't want to risk being "bumped up" as a result.

My question: I've heard it said that as a benchmark my rating will be determined through league play and then I will be "used" to help calculate others' ratings if I should play tournaments. Is this true? Will my tournament results affect my rating at all or is it already set in stone?

If these are sanctioned tournaments, then your results there will be factored into your year-end rating. From what I understand, 50% of your rating is determined by your championship matches and 50% from your local league matches. I'm guessing the tournament aspect gets factored in with the local league matches at the end of the year. So there is no immediate impact, but it could affect the likelihood you get bumped up to 4.5 at the year's end.

 

[HoustonHacker]

thanks. if 50 percent is local league matches, it seems as though the tournaments won't count, no?

 

[amarone]

If these are sanctioned tournaments, then your results there will be factored into your year-end rating. .

Unless it is different in other sections, NTRP tournament play does not count towards your year-end rating. From the Southern Section web site:

What data is not used to calculate year-end ratings?

Combo Doubles
Non USTA Leagues/Fun Leagues/Flex Leagues
NTRP Tournaments
Non-Sanctioned Tournaments
6-0, 6-0 match results

From what I understand, 50% of your rating is determined by your championship matches and 50% from your local league matches.

For non-benchmark players, the rating is made up 30% from their final dynamic rating and 70% from final dynamic with benchmarks factored in. For benchmark players, there is a statement on the usta.com site that more weighting is given to matches against other benchmark players, but this is not quantified anywhere that I can find.

 

[raiden031]

Unless it is different in other sections, NTRP tournament play does not count towards your year-end rating. From the Southern Section web site:



For non-benchmark players, the rating is made up 30% from their final dynamic rating and 70% from final dynamic with benchmarks factored in. For benchmark players, there is a statement on the usta.com site that more weighting is given to matches against other benchmark players, but this is not quantified anywhere that I can find.

Where did you get this 30/70 thing? I read 50/50 somewhere. Anyways tourneys count in some sections. Thats why they have Tournament exclusive (T) ratings.

 

[amarone]

Where did you get this 30/70 thing? I read 50/50 somewhere. Anyways tourneys count in some sections. Thats why they have Tournament exclusive (T) ratings.

The 30/70 was in a presentation called "NTRP Mythbusters" given by our District Director of Leagues. I still have the material and can confirm that it says 30/70. I think he said that the weighting had changed recently, so it could have been 50:50 previously, when you were told those numbers.

 

[HoustonHacker]

thanks amarone. so basically, tournament play is used for ratings calculations only for tournament exclusive people? so if someone plays two league matches and 30 tournament matches his/her rating is derived based on the league matches?

sorry if my question is ignorant...i'm not quite sure of the difference between the 30 percent portion and the 70 percent portion and what exactly is included in each.

 

[raiden031]

thanks amarone. so basically, tournament play is used for ratings calculations only for tournament exclusive people? so if someone plays two league matches and 30 tournament matches his/her rating is derived based on the league matches?

sorry if my question is ignorant...i'm not quite sure of the difference between the 30 percent portion and the 70 percent portion and what exactly is included in each.

Tournament Exclusive is a rating for people who only play tournaments and not leagues. League matches are factored into ratings on a daily basis. Tournament matches (only in certain sections) are factored into the rating only at the end of the year for year-end rating calculations.

 

[raiden031]

I think he said that the weighting had changed recently, so it could have been 50:50 previously, when you were told those numbers.

I remember the 50/50 ratio, but it was a long time ago that I read it on some website. I'm not sure if it was USTA or some other site. I never had a clear understanding of how year-end ratings are calculated and how benchmark matches are integrated with dynamic ratings. The only thing I ever really had a grasp on was how dynamic ratings are generated.

 

[HoustonHacker]

thanks...yeah, i knew what a tournament exclusive rating is. what i don't know, and hopefully amarone can clear this up, is: "what data goes into the 30 percent portion and what data goes into the 70 percent portion?"

 

[amarone]

thanks...yeah, i knew what a tournament exclusive rating is. what i don't know, and hopefully amarone can clear this up, is: "what data goes into the 30 percent portion and what data goes into the 70 percent portion?"

I believe it is the same data going into each, in terms of which matches count. I put the types of matches that don't count in the Southern Section in an earlier post. Maybe other sections are different, although that would seem strange - at Nationals you would then have teams competing against each other with ratings calculated using different methods.

The difference between the 30% and the 70% is that the latter is recalculated using benchmark players to normalize the ratings across the country. So at year end you may have a dynamic rating of 3.2 based on local matches. USTA recalculates that using the benchmarks to normalize (and I have never heard how that algorithm works) and let us say it comes up with 3.3. Your year-end rating is now (3.2 x 0.3) + (3.3 x 0.7) = 3.27.

 

[Jim A]

calculating a rating is akin to herding alligators, our district just had a long article on the ins and outs, but had no idea on the benchmark part of it

so amarone, what you are saying is that I played about 11 matches this year in USTA league play, with (1) against a benchmark

So my dynamic in those other 10 matches only comprise 30% of the weight, and 70% against a B who was playing his first single match for that portion of it all? My matches @ Districts had no benchmark players (well they will be next year)
http://bit.ly/4wNkpQ

My section does include Twilight (Fall) and Tournament play, however they don't factor them in until the end of year.

It's been fun trying to understand this computer, somewhere I think there is a short man behind a tall curtain in all this in a place called Oz

 

[amarone]

so amarone, what you are saying is that I played about 11 matches this year in USTA league play, with (1) against a benchmark

So my dynamic in those other 10 matches only comprise 30% of the weight, and 70% against a B who was playing his first single match for that portion of it all? My matches @ Districts had no benchmark players (well they will be next year)
http://bit.ly/4wNkpQ

No, that is not what I am trying to say. Your dynamic rating includes the match against the benchmark and there is no different weighting. This gives you your year-end dynamic rating.

Then the matches between benchmarks at Nationals, Sectionals and District Championships are used to calibrate across the country. For example, it may be discovered that 3.2s in NorCal are consistently beating 3.3s from Southern. This indicates that ratings have become out of step between those sections. The year-end process recalibrates this to try achieve parity such that, for example, a 3.2 in NorCal is the same as a 3.2 in Southern. They start with Nationals and look at the results there. Note that when I say "they", this is actually done by computer algorithm - there is no person saying "Hmm, looks like NorCal is out of step." The algorithm then adjusts the ratings of these benchmarks across the sections based on results at Nationals. This is then repeated across the results at each sectional to recalibrate across the districts. Then repated at District to calibrate across the local leagues. Having done all this recalibration, your season's results are recalculated to give the rating that contributes 70% of your final rating. I hope that is clear.

 

[CrocodileRock]

Thanks for sharing how the system works amarone, in this thread and others. I just have one question:

Note that when I say "they", this is actually done by computer algorithm - there is no person saying "Hmm, looks like NorCal is out of step."

Are there some human beings in some of the sections or each section that do say "Hmmm. Player A is rated higher than Player B, even though B is clearly stronger. Let's fix that."

I think the system works pretty well overall, but we all probably know misrated players, which can be pretty frustrating.

 

[couch]

Why not just play the tournaments with your buddy and not worry about it? Why are you so worried about getting bumped? Go play tennis and have fun.

 

[Jim A]

much appreciated amarone

 

[10sguy]

Why not just play the tournaments with your buddy and not worry about it? Why are you so worried about getting bumped? Go play tennis and have fun.

Amen to that!!

 

[HoustonHacker]

thanks amarone. it's not just about playing the tournaments with my buddy, i've also been curious about how the end of the year ratings are calculated, as there is little written about it and a lot written about in-season dntrp calculations for red flag purposes. amarone shed some light on the subject for me, and as a math geek i find it fascinating.

plus, bill, i'm actually trying to "figure out" whether i'll be bumped up based on my sectionals results, etc. so i can plan for next year in advance. hopefully i'll see you at 4.0 sectionals again...that was quite a match you guys played against us and you should be proud.