PDA

View Full Version : Rating statistics


raiden031
12-13-2008, 09:17 AM
I was playing around and put all the new ratings of players in my local county into an excel spreadsheet. I then calculated statistics at the various levels to see what the distribution is. For those who might be interested, this is what I found. I'm not sure if my county is a good sample but I figure it might be useful to some.

Rated Players: 1548 (902 F, 646 M)

Male Statistics:

2.0 - 001 - 00.15% (Top 100%)
2.5 - 003 - 00.46% (Top 99%)
3.0 - 128 - 19.81% (Top 80% - 99%)
3.5 - 282 - 43.65% (Top 36% - 80%)
4.0 - 171 - 26.47% (Top 9% - 36%)
4.5 - 053 - 08.20% (Top 1% - 9%)
5.0 - 008 - 01.24% (Top 0% - 1%)

Female Statistics:

2.5 - 119 - 13.19% (Top 87% - 100%)
3.0 - 306 - 33.92% (Top 53% - 87%)
3.5 - 338 - 37.47% (Top 15% - 53%)
4.0 - 106 - 11.75% (Top 4% - 15%)
4.5 - 033 - 03.66% (Top 0% - 4%)

Nellie
12-13-2008, 09:56 AM
I have felt for a while that the USTA was pushing players toward 3.5/4.0s, with the 3.0 being beginner players and 4.5's being very experienced players used to competition.

OrangePower
12-13-2008, 03:57 PM
That's interesting. First impressions:

1. Makes you wonder why USTA bothers to have 2.0 and 2.5 levels for men. I mean, a total of 4 out of 646 in these levels?!? What's the point?

2. I didn't realize how much the distributions differ between females and males. I wonder why. The USTA states that ratings are relative only within gender (i.e., a 3.5 male and 3.5 female do not have the same absolute skill level). So I always assumed that the intent is to have the distribution curves be similar for each gender, but these numbers disprove that assumption. I now have more respect for female 4.0s... reaching 4.0 as a female is about as hard as reaching 4.5 as a male.

3. Raiden has way to much free time on his hands :)

goober
12-13-2008, 04:30 PM
That's interesting. First impressions:



2. I didn't realize how much the distributions differ between females and males. I wonder why. The USTA states that ratings are relative only within gender (i.e., a 3.5 male and 3.5 female do not have the same absolute skill level). So I always assumed that the intent is to have the distribution curves be similar for each gender, but these numbers disprove that assumption. I now have more respect for female 4.0s... reaching 4.0 as a female is about as hard as reaching 4.5 as a male.:)


It depends. I don't think you can automatically draw that conclusion based on the percentage of 4.5 and 4.0 men and women. If your population base of women players in your section are mostly seniors or social club players, it may not really be harder to get to 4.0. Since the formula for dynamic ratings are the same for both genders, it demonstrates in my view that there is a higher percentage of male players that become dominant at a specific level that moves them up.

OrangePower
12-13-2008, 07:30 PM
It depends. I don't think you can automatically draw that conclusion based on the percentage of 4.5 and 4.0 men and women. If your population base of women players in your section are mostly seniors or social club players, it may not really be harder to get to 4.0. Since the formula for dynamic ratings are the same for both genders, it demonstrates in my view that there is a higher percentage of male players that become dominant at a specific level that moves them up.

You're right, we don't know what the specific demographic makeup of tennis players is in this county (stats are for Raiden's area, not mine).

But the bottom line is that in this area, for the women, reaching 4.0 means being in the top 10% or so, while for men, in order to be in the top 10%, you need to get to 4.5. Ergo, in this area, a 4.0 woman has reached the same relative level of accomplishment as a 4.5 man.

Coach Carter
12-13-2008, 07:59 PM
I have posted in some other sections about this. There are very few "rated" players over 4.5 - BECAUSE there isn't much competition there in most areas. If you want to play, then you'll self rate and play down...the sections aren't going to move you out of the norm in your area unless there is something glaring. In my area, there are some real good guys playing 4.0 and 4.5...because they want to play...it's all good.

goober
12-14-2008, 05:20 AM
You're right, we don't know what the specific demographic makeup of tennis players is in this county (stats are for Raiden's area, not mine).

But the bottom line is that in this area, for the women, reaching 4.0 means being in the top 10% or so, while for men, in order to be in the top 10%, you need to get to 4.5. Ergo, in this area, a 4.0 woman has reached the same relative level of accomplishment as a 4.5 man.

Yes but relative accomplishment is meaningless when you are comparing 2 dissimilar groups. In colloquial terms it is an apples and oranges comparison. Let's take men and women of out the equation and use an comparison which by design is somewhat extreme.

Group A: top 10% of runners at 100 m from a junior high phys ed class

Group B: top 10% of runners at 100 m from a group of Olympic sprinters.


While the relative level accomplishment is the same for both groups, it is a meaningless comparison. Just being in group B is an accomplishment that very few if any in group A could hope to achieve.

So interesting numbers, but I don't have a new found respect or disrespect for a a specific subgroup. For me absolute difficulty to reach a certain level accomplishment is much more important than relative difficulty within subgroup.

end of rant- carry on :mrgreen:

raiden031
12-14-2008, 03:25 PM
3. Raiden has way to much free time on his hands


It didn't take that long to do, and I was curious enough to want to put in the effort.

You're right, we don't know what the specific demographic makeup of tennis players is in this county (stats are for Raiden's area, not mine).

But the bottom line is that in this area, for the women, reaching 4.0 means being in the top 10% or so, while for men, in order to be in the top 10%, you need to get to 4.5. Ergo, in this area, a 4.0 woman has reached the same relative level of accomplishment as a 4.5 man.

I might look at another area and see if it matches these stats.

What I find interesting is that you could easily revert into the old way of having 3 different skill divisions instead of 6 (2.5-5.0). Why create skill divisions for like 10% of players to be divided up into?

For me absolute difficulty to reach a certain level accomplishment is much more important than relative difficulty within subgroup.

I think if you are in a popular tennis area than these numbers might mean something (I'm more interested in the frequency distribution of NTRP levels for my gender rather than comparing the genders against each other). I think my area is reasonably strong and its nice to see where I fit in compared to my peers. I should check an area like Northern VA that has double the amount of players and see what their stats look like.

randomname
12-14-2008, 03:30 PM
hmm, I wonder what the reason for more men being closer to 4.0 and more women being closer to 3.0 is. It doesnt seem like it would be men practice more because I know more female players who dont work and have time during the day to go play than men. Perhaps its because more women start playing later in life than men?

raiden031
12-14-2008, 03:39 PM
hmm, I wonder what the reason for more men being closer to 4.0 and more women being closer to 3.0 is. It doesnt seem like it would be men practice more because I know more female players who dont work and have time during the day to go play than men. Perhaps its because more women start playing later in life than men?

Probably because more men take up sports as kids. Many women players came into tennis as adults without playing prior sports growing up.

I'm going to post stats from another area in an hour or two before speculating too much based on my area's stats.

JavierLW
12-14-2008, 05:03 PM
hmm, I wonder what the reason for more men being closer to 4.0 and more women being closer to 3.0 is. It doesnt seem like it would be men practice more because I know more female players who dont work and have time during the day to go play than men. Perhaps its because more women start playing later in life than men?

Right, I think it's precisely to that point.

Among Adults who didnt play as much when they were younger (which usually accounts for a lot of the higher level players), and have just picked up the game now, there are a LOT more women players then men.

I think especially in winter areas where most winter tennis is expensive indoor tennis. I know at our clubs women's adult events are typically huge, and with the men (which I am pretty active in), it's hard to get the same people to show up from week to week.

Also if you look at the numbers in our leagues at least in our area (check out any winter area sometimes), the women's teams outnumber the men's teams sometimes by as much as 4-1.

So I dont think it has anything to do with anyone ability to improve, it just means there are more women involved in that particular nitch of tennis then there are men.

There are more women entering the sport as well which accounts for the higher 3.0 numbers. If their numbers were smaller, you'd probably see a similar pattern then what happens with men (most of them move up to 3.5).

OrangePower
12-14-2008, 07:34 PM
It didn't take that long to do, and I was curious enough to want to put in the effort.

Hey Raiden, I hope you didn't take my comment as a dig; it was made tongue in cheek and I actually think it's cool you posted this (I'm kinda a stats nerd).

What I find interesting is that you could easily revert into the old way of having 3 different skill divisions instead of 6 (2.5-5.0). Why create skill divisions for like 10% of players to be divided up into?

Definitely agree.

raiden031
12-15-2008, 07:50 AM
Mid-Atlantic - NOVA

Rated Players 3172 (1468M / 1704F)

Male Statistics:

2.5 - 005 - 00.34% (Top >99%)
3.0 - 201 - 13.69% (Top 86% - >99%)
3.5 - 500 - 34.06% (Top 52% - 86%)
4.0 - 500 - 34.06% (Top 18% - 52%)
4.5 - 237 - 16.14% (Top 2% - 18%)
5.0 - 022 - 01.50% (Top <1% - 2%)
5.5 - 001 - 00.07% (Top <1%)
6.0 - 001 - 00.07% (Top <1%)
6.5 - 001 - 00.07% (Top <1%)

Female Statistics:

2.5 - 034 - 02.00% (Top >98%)
3.0 - 476 - 27.93% (Top 70% - 98%)
3.5 - 654 - 38.38% (Top 32% - 70%)
4.0 - 399 - 23.42% (Top 8% - 32%)
4.5 - 134 - 07.86% (Top <1% - 8%)
5.0 - 007 - 00.41% (Top <1%)

------
Stats within the stronger Northern VA area. Amazing that 50% of male players are at least a 4.0. The women are still lagging behind but are more in line with the men in my county (Howard, MD).

DBH
12-15-2008, 09:51 AM
From raiden's most recent statistics, it looks as if there are, for all intents and purposes, four levels: 3.0, 3.5, 4.0, and 4.5.

If you're lower than 3.0, you're probably better off practicing and playing friendly matches to gain experience than USTA/league matches. If you're higher than 4.5, you're probably better off playing open tournaments than USTA/league matches.

Of course, there are always a few exceptions, but it looks like from a practical point of view, almost all league tennis players fall within four levels. Which seems fine to me -- the idea is to maintain competitive matches within levels, while not spreading the playing opportunities too thin by having too many finely defined levels.

Any comments as to whether the current system is successful in doing that?

JavierLW
12-15-2008, 10:00 AM
From raiden's most recent statistics, it looks as if there are, for all intents and purposes, four levels: 3.0, 3.5, 4.0, and 4.5.

If you're lower than 3.0, you're probably better off practicing and playing friendly matches to gain experience than USTA/league matches. If you're higher than 4.5, you're probably better off playing open tournaments than USTA/league matches.

Of course, there are always a few exceptions, but it looks like from a practical point of view, almost all league tennis players fall within four levels. Which seems fine to me -- the idea is to maintain competitive matches within levels, while not spreading the playing opportunities too thin by having too many finely defined levels.

Any comments as to whether the current system is successful in doing that?

I think the current system reflects that there are differences between a 4.0 a 4.5, a 5.0 and even a 5.5 or a 6.0 (open level)

So moving all of those higher level players into a lower level is not always ideal even if it means there are more teams because those players will clearly win and the matches will not necessarily be competitive.

This also goes with the idea that the USTA uses that it's usually not a big deal if you play up (most people get better by playing up anyway), but it is a big deal if you play down, because you will win for sure.

Also even though a local league may have this sort of participation, if you look at from a national level, they have playoffs for all of these levels with about the same participation. And that seems to be where the USTA's focus is most of the time (unfortuanlly).

If they have a 2.5 National or a 5.5 National they are not even thinking about what's going on in your local league.

I dont think it's a big deal now. Most areas have a 3.0, a 3.5 and a 4.0, and usually a 4.5.

I dont think it matters if they have higher levels but moving all of those players into 4.5 is not a solution. I think they belong at a different level and it doesnt hurt anything if there is not a lot of participation at that level.

Having those players just concentrate more on tournaments is really the answer (and the reason why the numbers are as few as they are, besides that there are less of those players period).

(also the USTA does not have a 2.0 level anywhere as far as I know. Some players have a 2.0 rating, but that usually means that either they self rated there or maybe they did really bad against 2.5 level players. I feel the reason why there are even 2.5 level players at all in some cases is because the REAL 3.5 players were all playing in the 3.0 league so the REAL 3.0 players got rated down to 2.5. There really "shouldnt" be as many 3.0 players as there are, it's just that everyone starts out there even though a lot of people could of probably started in 3.5 because with any amount of talent you can skip over 3.0 just like 2.0 and 2.5, it's not quite that easy to get out of 3.5 to 4.0)

Morpheus
12-15-2008, 11:44 AM
Looks like there's an awful lot of bad tennis being played.

I for one wish they would collapse the 4.0/4.5 into one group because the numbers of 4.5 are low and it is hard to fill a league in a lot of regions. This doesn't seem to matter as much in the 3.5 range where there are a lot of numbers.

raiden031
12-15-2008, 01:05 PM
Looks like there's an awful lot of bad tennis being played.

I for one wish they would collapse the 4.0/4.5 into one group because the numbers of 4.5 are low and it is hard to fill a league in a lot of regions. This doesn't seem to matter as much in the 3.5 range where there are a lot of numbers.

On the surface it does seem silly to have like 6 levels when most of the players are concentrated in around 3 of them. But I guess the problem is that due to the nature of the rating algorithm, it wouldn't be feasible to support less levels with a wider range of skill levels. There would be far too many "meaningless" matches (those that are 6-0, 6-0 or 6-0, 6-1 matches) where it can't really be determined how close the dynamic ratings of the players actually are.

raiden031
12-15-2008, 01:17 PM
(also the USTA does not have a 2.0 level anywhere as far as I know. Some players have a 2.0 rating, but that usually means that either they self rated there or maybe they did really bad against 2.5 level players. I feel the reason why there are even 2.5 level players at all in some cases is because the REAL 3.5 players were all playing in the 3.0 league so the REAL 3.0 players got rated down to 2.5. There really "shouldnt" be as many 3.0 players as there are, it's just that everyone starts out there even though a lot of people could of probably started in 3.5 because with any amount of talent you can skip over 3.0 just like 2.0 and 2.5, it's not quite that easy to get out of 3.5 to 4.0)

Anything below 2.5 is pointless. I think 2.5 is a special case in that they should only have a local league and thats it. Having a 2.5 Nationals is a complete joke because people can advance past 2.5 very quickly.

I couldn't believe how difficult it was for me to get bumped out of 3.0 (back in '07 where I stayed at 3.0). People often associate 3.0s with beginner tennis, yet I was hitting kick serves and pounding topspin shots, yet losing to people enough at that level to stay there. Nobody ever would have said I was a beginner back then. All the tough players seemed to be self-rated too.

JavierLW
12-15-2008, 02:21 PM
Anything below 2.5 is pointless. I think 2.5 is a special case in that they should only have a local league and thats it. Having a 2.5 Nationals is a complete joke because people can advance past 2.5 very quickly.

I couldn't believe how difficult it was for me to get bumped out of 3.0 (back in '07 where I stayed at 3.0). People often associate 3.0s with beginner tennis, yet I was hitting kick serves and pounding topspin shots, yet losing to people enough at that level to stay there. Nobody ever would have said I was a beginner back then. All the tough players seemed to be self-rated too.

Right, I think 2.0 and 2.5 are a waste of time as well. Especially 2.0.

I think if you play a friend or someone who simply never plays tennis maybe you could say they are a 2.0 or a 2.5, but if they did play tennis with any frequency they would at least end up in 3.0, especially if they are good at some other sports.

The problem is that those are skill ratings, and if some teaching pro had to evaluate someone there are people who probably are 3.0's. But when you look at the League, the ratings are polluted because like you said, many of the players are self rated and probably did not belong there.

Then they go on to play other players that are similar and their rating gets stuck at 3.0. (but a few of them enjoy that because they can move on and win a pen or something in the playoffs)

That's why at some point I stopped playing 3.0 all together even though I still had a 3.0 rating and I only had a 42% win rate there. I just figured that among the best players in 3.0, most of them are 3.5's (or better) anyway. And 3.5 has plenty of weaker players (which are the same average players you see in 3.0) as well.

Other than meeting the 4.0ish teams, there really wasnt much of a difference.

Ive seen plenty of people do this. When they play both 3.0 and 3.5, their rating gets dragged down to 3.0, but when they stop playing 3.0, their win rate remains about the same, and they get rated up to 3.5.

David_Is_Right
01-26-2009, 01:56 PM
I really think the sample size is too small and the background information on the sample is too sparse to draw general conclusions here.

raiden031
01-26-2009, 02:55 PM
I really think the sample size is too small and the background information on the sample is too sparse to draw general conclusions here.

Right, thats why I put a bit of a disclaimer in there. But one thing that is true is that these areas are rich in tennis programs for adult rec. players though.

J-Mac
01-31-2009, 06:29 AM
Interesting stats

It would also be interesting to see which groups play the most tennis.

It's been my experience that 3.5 players play the most per week. I think they are still improving their games and it's very realistic for them to become 4.0 players. It's a lot more encouraging to play when you still see progress in your game.

4.0 players don't play as often because they have pretty much reached their peak other than a few that will move into the 4.5/5.0 range.


______________________
http://www.tennistour.org
Tennis Everyday - All Skills

Slicendicer
01-31-2009, 07:04 AM
I have posted in some other sections about this. There are very few "rated" players over 4.5 - BECAUSE there isn't much competition there in most areas. If you want to play, then you'll self rate and play down...the sections aren't going to move you out of the norm in your area unless there is something glaring. In my area, there are some real good guys playing 4.0 and 4.5...because they want to play...it's all good.

A couple of years ago while sub-ing in a local 4.5 tennis league, I played singles with a guy that a very good player. After the match, he said 2 years ago he played 1 singles at Rice and was currently playing challengers. There is no 5.0 league, the 4.5's are majority 5.0's and current D1 college players.

dabudabuda
01-31-2009, 11:01 AM
with the exception of the mid-atlantic 5.5+(outliers/influential data...) men, most sets of data are fairly normal/bell curved. the women are close as well maybe just slighty skewed but if you take out the highest rating quite "normal".

if you could somehow break down the ratings even further, it would look more normal on both sides.

cak
01-31-2009, 06:39 PM
Okay, from the Norcal site I pulled South Bay, Lower Pen, and Mid Pen (pretty much Santa Clara County)

Rated Players: 6077 (2664 F, 3413 M)

Male Statistics:

2.0 - 0 - 00.00%
2.5 - 46 - 01.35%
3.0 - 592 - 17.34%
3.5 - 1331- 39.00%
4.0 - 938- 27.48%
4.5 - 386- 11.31%
5.0 - 96- 02.81%
5.5 - 24- 00.70%

Female Statistics:
2.0 - 2 - 00.07%
2.5 - 165 - 06.19%
3.0 - 759 - 28.49%
3.5 - 1107 - 41.55%
4.0 - 500 - 18.78%
4.5 - 111 - 04.17%
5.0 - 17 - 00.64%
5.5 - 3 - 00.11%

For men, about 58% are 3.5 or below, and the median is somewhere in the 3.5 range. For women 76% are 3.5 or below, but the median is still in the 3.5 range. I'm not sure about men, but for the women in the last two years sweet spot for women seems to have moved from 3.0 to 3.5, in that a few years ago there were more 3.0 women than 3.5 women.

(And for those playing on the NorCal site, the report that runs this doesn't double count if you play in more than one district.)