NTRP Questions...

[COLuv2Play]

leagueCO -- I have some questions

I am in Colorado where everything seems to count -- all of my league participation matches (USTA leagues and CTA leagues) and all of my sanctioned tournament matches. Looking at my record for 2012 so far this year, I believe I am really close to moving up from 3.0 to 3.5. I wondered how much the Benchmark matches figure in to this and how USTA looks at a player who wins over 3.5 benchmark players (sometimes) but loses to 3.0 benchmark players (sometimes). Is it the same kind of math that is applied to any other match, just with maybe a multiplier of 2 or something?

I want to do everything I can in the little time that is left to make sure I move up. I basically have decided that playing against 3.0 players whether I win or lose is a lot riskier for me than playing 3.5 -- especially if I am pretty sure I can post a competitive score at 3.5.

Your thoughts?

[Orange]

I cannot point you to a particular source, but I have researched NTRP using every source on the internet because I captained 3 teams this year, and I do not believe that matches against benchmark players count any more or less than matches against other players. A benchmark rating is one based on playoff results.

It is possible to beat 3.5 benchmark players and lose to 3.0 benchmark players because those benchmark ratings are based on results from last year. A 3.0 player who is rapidly improving could have a higher dynamic rating that is not reflected in his 3.0 rating that is visible to the public.

[COLuv2Play]

Orange -- I don't lose to 3.0 Benchmark players because they are better than 3.5 Benchmark players. They definitely are not! I lose to them because I don't play as well. The better my opponents are, the better I play.

The question really has nothing to do with why I lose to 3.0s and win over 3.5s, it has to do with how the USTA views that.

[Orange]

I'm sorry I misunderstood your question.

Also, I should have said that benchmarks are based on post-season play, not playoffs.

[tennis tom]

Originally Posted by COLuv2Play

leagueCO -- I have some questions

I don't think LeagueCO is in any hurry to come back here, if ever. I may try to contact him to ask him to return if he left an email address.

[schmke]

Originally Posted by Orange

I cannot point you to a particular source, but I have researched NTRP using every source on the internet because I captained 3 teams this year, and I do not believe that matches against benchmark players count any more or less than matches against other players. A benchmark rating is one based on playoff results.

It is possible to beat 3.5 benchmark players and lose to 3.0 benchmark players because those benchmark ratings are based on results from last year. A 3.0 player who is rapidly improving could have a higher dynamic rating that is not reflected in his 3.0 rating that is visible to the public.

It isn't so much that benchmark ratings count more than other results, but that once the benchmark players are identified and their ratings calculated, recalculations are done that could change your and other players ratings from just the regular Dynamic NTRP that is calculated throughout the year. I've seen mentions that this benchmark calculation may be 50% of the year end rating with the other 50% the normal Dynamic NTRP.

[Mongolmike]

Originally Posted by Orange

I cannot point you to a particular source, but I have researched NTRP using every source on the internet because I captained 3 teams this year, and I do not believe that matches against benchmark players count any more or less than matches against other players. A benchmark rating is one based on playoff results.

My understanding is by virtue of being a benchmarked player- (which means they have advanced in post season matches, right?) - that in itself raises their rating because they are at the top of the pyramid for their rating... so if you've played against a benchmark (win or lose) that factors into the computer system as a tougher opponent.

In other words, many players who advance in post season will get benchmarked and they also have a higher chance of being bumped up... they and their team have shown by being in the post season (to a certain degree) that they are the best at their rating.... so if you've played against a benchmarked player, that figures in the computation as a tougher match, win or lose.

During the season, it doesn't matter as much because the opponent is not benchmarked yet... but at some point during their playoff run they will be benchmarked, and at that point that changes the factors into how it affects your rating/your match vs them months earlier... right?

[dizzlmcwizzl]

Originally Posted by Mongolmike

My understanding is by virtue of being a benchmarked player- (which means they have advanced in post season matches, right?) - that in itself raises their rating because they are at the top of the pyramid for their rating... so if you've played against a benchmark (win or lose) that factors into the computer system as a tougher opponent.

In other words, many players who advance in post season will get benchmarked and they also have a higher chance of being bumped up... they and their team have shown by being in the post season (to a certain degree) that they are the best at their rating.... so if you've played against a benchmarked player, that figures in the computation as a tougher match, win or lose.

During the season, it doesn't matter as much because the opponent is not benchmarked yet... but at some point during their playoff run they will be benchmarked, and at that point that changes the factors into how it affects your rating/your match vs them months earlier... right?

NO, NO, NO .... Being bench marked itself has nothing to do with your rating. It just means you played in the postseason.

However, if you played in the postseason, the chances are greater that you are a good player and at the top of your band ... otherwise you would have not been in the lineup presumably. But this is not always the case for obvious reasons.

None of us actually know the algorithm but this is my conjecture as to why the USTA cares at all about the benchmarks ...

I played in sectionals for Middles States at 4.0. When I played our section's national representative I lost in doubles in a third set match tiebreaker. This was considered a tie because even though they won the match we split games.

Now if the guys we played go out to nationals and get smoked then the USTA will say "Dizzl, yall are about the same as these guys and they got smoked so we are gonna dial you back" (and other guys that played against these dudes) to keep the sections as equal as possible. However, if our guys go out and dominate Nationals, the USTA will raise us a little and lower some other sections.

Being benchmarked does not make you more likely to get bumped, performing better than your peers is what drives the bump...

In fact, last year in our district 10 folks were early start bumped. All were benchmarked and after the team from our section performed poorly at nationals every single one was moved back down at year end. However, 3 players that were not benchmarked were moved up at year end even though they were not early start bumps.

[kylebarendrick]

Having a benchmark rating does NOT mean that someone is at the top of their level. If someone plays at Districts as a 3.5 and then gets bumped at the end of the year. Their new rating will be 4.0b - even though they have never played a match at 4.0.

Anyone with a 4.0b is either a 3.5 that played in the post season and got bumped or a 4.0 that played in the post season and didn't get bumped. Those are the only conclusions that you can draw.

[Mongolmike]

Originally Posted by dizzlmcwizzl

NO, NO, NO .... Being bench marked itself has nothing to do with your rating. It just means you played in the postseason.

However, if you played in the postseason, the chances are greater that you are a good player and at the top of your band ... otherwise you would have not been in the lineup presumably. But this is not always the case for obvious reasons.

Ok, you said it a bit differently... the point is the same tho...

-IF you play in the post season you will get benchmarked.

-IF you play in the post season you are presumably a good player and near the top of your band. You will be benchmarked AND your dynamic rating will improve as you advance. Your dynamic rating can change match by match even.

-IF I played a player in the regular season who later played successfully in the post season, this will have an upward effect on my end of year rating.

So you can get a benchmark if you do well in the post season... so its not the benchmark that effects ratings, but the fact you advanced in the post season.

Post season matches=change in your dynamic rating and a benchmark= change in your regular season opponent's end of year algorithm

[dizzlmcwizzl]

Originally Posted by Mongolmike

Ok, you said it a bit differently... the point is the same tho...

-IF you play in the post season you will get benchmarked.

-IF you play in the post season you are presumably a good player and near the top of your band. You will be benchmarked AND your dynamic rating will improve as you advance. Your dynamic rating can change match by match even.

-IF I played a player in the regular season who later played successfully in the post season, this will have an upward effect on my end of year rating.

So you can get a benchmark if you do well in the post season... so its not the benchmark that effects ratings, but the fact you advanced in the post season.

Post season matches=change in your dynamic rating and a benchmark= change in your regular season opponent's end of year algorithm

Just to be clear ... playing well in the postseason did not necessarily "help" my rating. Last year during the regular adult season I went 19-1 and then in district and sectionals I went a combined 7-2. I was early start rated up to 4.5 at the end of the season. However, one of the the two teams I lost to at sectionals got smashed at nationals ... their results moved me back down. Of course I don't know, but I suspect if I had not played at sectionals last year I would have been moved up at year end. Therefore playing in the postseason actually likely "hurt" my rating.

[Islandtennis]

One thing to remember is that the primary use of benchmarks is to equate NTRP levels across local areas. The district, sectional, and national playoffs is the only time (in league play) that people play out of their local area. It attempts to do the impossible of making a 3.5 in CA similar to a 3.5 in NY; or a 3.5 in one part of GA similar to a 3.5 in another part of GA. While not perfect, it does work better than having no equalizing in place.

[dizzlmcwizzl]

Originally Posted by kylebarendrick

Having a benchmark rating does NOT mean that someone is at the top of their level. If someone plays at Districts as a 3.5 and then gets bumped at the end of the year. Their new rating will be 4.0b - even though they have never played a match at 4.0.

Anyone with a 4.0b is either a 3.5 that played in the post season and got bumped or a 4.0 that played in the post season and didn't get bumped. Those are the only conclusions that you can draw.

It could be a 4.5 player that played in playoffs and then was moved down. Seems implausible but I know of exactly this case ... where a player played 4.5 districts this year and was bumped down at early start rating time.

[Spokewench]

This is how I understand Benchmark - it means if you go on to advance to post season play; say you have a dynamic rating of 3.466 you are a benchmark for that dynamic rating. Not, that you are at the height of your rating level, not that you are at the low end. You set the benchmark for that particular dynamic rating.

I have been to sectionals several times, but I am also from a very small district so I have also gotten my butt kicked each and every time I have gone to sectionals (i was not at the top of the rating level).

[ian2]

Originally Posted by Islandtennis

One thing to remember is that the primary use of benchmarks is to equate NTRP levels across local areas. The district, sectional, and national playoffs is the only time (in league play) that people play out of their local area. It attempts to do the impossible of making a 3.5 in CA similar to a 3.5 in NY; or a 3.5 in one part of GA similar to a 3.5 in another part of GA. While not perfect, it does work better than having no equalizing in place.

Great discussion! My thoughts are very much in-line with what dizzlmcwizzl's was saying, and I agree with the above as well. But to me, "not perfect" is the key here: I feel that NTRP's attempt to "equate levels across local areas" is based on a faulty premise. A set of faulty premises actually: the algorithm makes statistical determinations based on a sample size that is too small to support such determinations; on top of that, the sample itself is likely to be NOT representative of the "average" level of play in a given area. Then the problem is compounded by giving this questionable "denominator" too much weight.

I'm actually not sure that this works better than having no equalizing in place, even for the stated purpose of equating NTRP levels across local areas. As for the role this plays in calculating individual year-end DNTRP: while benchmark recalculation produces a ripple effect for all players, it almost certainly affects players who happened to play directly against benchmarks in a non-linear way, and can result in some odd NTRP outcomes for such players. It appears that one blow-out loss (or win... it goes both ways) against a benchmark player might outweigh a player's entire record during the year.

[Orange]

Originally Posted by dizzlmcwizzl

It could be a 4.5 player that played in playoffs and then was moved down. Seems implausible but I know of exactly this case ... where a player played 4.5 districts this year and was bumped down at early start rating time.

This was the case with a player on my team. She was bumped down at year-end (not early start) after losing badly in the playoffs last summer. I am, of course, thrilled to have her on my team.

[J_R_B]

Originally Posted by ian2

Great discussion! My thoughts are very much in-line with what dizzlmcwizzl's was saying, and I agree with the above as well. But to me, "not perfect" is the key here: I feel that NTRP's attempt to "equate levels across local areas" is based on a faulty premise. A set of faulty premises actually: the algorithm makes statistical determinations based on a sample size that is too small to support such determinations; on top of that, the sample itself is likely to be NOT representative of the "average" level of play in a given area. Then the problem is compounded by giving this questionable "denominator" too much weight.

I'm actually not sure that this works better than having no equalizing in place, even for the stated purpose of equating NTRP levels across local areas. As for the role this plays in calculating individual year-end DNTRP: while benchmark recalculation produces a ripple effect for all players, it almost certainly affects players who happened to play directly against benchmarks in a non-linear way, and can result in some odd NTRP outcomes for such players. It appears that one blow-out loss (or win... it goes both ways) against a benchmark player might outweigh a player's entire record during the year.

If you didn't at least TRY to normalize the ratings across areas, then you would definitely get the ratings "drift" that people constantly claim on here in stronger tennis areas but really doesn't actually exist and a bias towards larger areas at nationals. The normalization isn't perfect for sure, but it does do as good of a job as it can of at least eliminating the creation of hot spots whose teams win nationals every year.

[floridatennisdude]

Originally Posted by J_R_B

If you didn't at least TRY to normalize the ratings across areas, then you would definitely get the ratings "drift" that people constantly claim on here in stronger tennis areas but really doesn't actually exist and a bias towards larger areas at nationals. The normalization isn't perfect for sure, but it does do as good of a job as it can of at least eliminating the creation of hot spots whose teams win nationals every year.

Couldn't agree more. I live in a tennis "hotbed" and we have had teams advance to, and win, nationals over the years. However, our chances of having a team accomplish this are only increased by volume of competitors. We don't get a team to nationals every year, but we are in the hunt at sectionals because captains have tons of players to choose from.

Overall though, we are pretty equal from a talent perspective with any other area of the country. Not like we can take our local league champ on a tour of the country and dominate city-by-city. That's why I see NTRP being reasonably effective at providing balanced competition. Not perfect, but pretty good.

I still feel that self rates should be ineligible for post season. They need a C to confirm their level to keep it the most fair it can be.

[schmke]

Originally Posted by J_R_B

If you didn't at least TRY to normalize the ratings across areas, then you would definitely get the ratings "drift" that people constantly claim on here in stronger tennis areas but really doesn't actually exist and a bias towards larger areas at nationals. The normalization isn't perfect for sure, but it does do as good of a job as it can of at least eliminating the creation of hot spots whose teams win nationals every year.

Originally Posted by floridatennisdude

Couldn't agree more. I live in a tennis "hotbed" and we have had teams advance to, and win, nationals over the years. However, our chances of having a team accomplish this are only increased by volume of competitors. We don't get a team to nationals every year, but we are in the hunt at sectionals because captains have tons of players to choose from.

Overall though, we are pretty equal from a talent perspective with any other area of the country. Not like we can take our local league champ on a tour of the country and dominate city-by-city. That's why I see NTRP being reasonably effective at providing balanced competition. Not perfect, but pretty good.

I still feel that self rates should be ineligible for post season. They need a C to confirm their level to keep it the most fair it can be.

Agreed as well. We all know the NTRP isn't perfect and some people manipulate it or take advantage of how it works, but it is better than nothing and trying to normalize across sections is certainly better than not trying. Could it be improved? Sure, but let's not just throw it out or not try to normalize.

I also agree that self-rates are an issue, but not because they are allowed to play post-season, but more because the thresholds the USTA sets for DQs are too high and/or that they don't DQ a player after a certain point. I have seen numerous cases where my estimated DNTRP has indicated a player has 3+ results in excess of what I'd expect the threshold to be, but they don't get DQ'd. It appears, at least at the 3.5 level where I've observed this primarily, but perhaps other levels too, the room the USTA gives for "natural improvement" is too large and allows self-rates who are clearly above level to continue to play at a level.

http://computerratings.blogspot.com/search/label/tennis

[floridatennisdude]

^^^great post

My opinion on self rates not playing post season is simple. Most people join a competitive league to win. Very few would be psychologically capable of ranking matches for an entire year, just to possibly dominate in year 2. But, if a few sneak through that way...all the power to them.

I just think you'll end up with far fewer sandbaggers in year 2 than year 1 if you get a year of data on them. To me, it just seems simple. I don't care if self rates are in a local league, but soooo many get DQd at districts and sectionals that it makes me wonder...what team would've been here that played by the rules instead of this team that cheated?