PDA

View Full Version : 2010 ratings statistics (Denver Metro)


ian2
11-30-2010, 09:34 AM
I run some stats on the entire Denver Metro area men Adult leagues, here is the summary:

2.5 total - 35 / moved up - 21 (60.0%)
3.0 total - 354 / moved up - 47 (13.3%; 3 players moved from 3.0 to 4.0)
3.5 total - 1085 / moved up - 56 (5.2%; 1 player moved from 3.5 to 4.5)
4.0 total - 1167 / moved up - 75 (6.4%)
4.5 total - 456 / moved up - 13 (2.9%; 1 player moved from 4.5 to 5.5)
5.0 total - 116 / moved up - 6 (5.2%)

The stats are probably a little off because some players played up in a league and were bumped to match the league's level. I.e. some of the year-end 3.5 players who played 3.5 league were actually 3.0 at the start of the year. This would account for no more than a couple of percentage points though, and likely substantially less than that.

There was no movement down to speak of... it's a bit harder to determine but I'm certain that no more than 1% of all players were moved down, at any NTRP level.

Last year as you know USTA manipulated the NTRP algorithm to generate a big bump-up across the board... for example, 30% of all 3.5 players nationally were moved up; even higher percentages for 3.0 and 2.5. It looks like this year the pendulum swung the other way. I wonder whether 5-6% up-movement for 3.5 - 5.0 levels is within the "normal" (pre-2009) range? It looks like this year USTA tweaked the algorithm to nearly eliminate any movement (except at 2.5 and to a lesser extent 3.0).

polski
11-30-2010, 10:20 AM
Those numbers look pretty normal and expected to me. Seeing as players generally increase in ability over time & rarely decrease in ability it shouldn't be expected to see much in terms of downward movement.

Players that may seem like they should move down are usually OK at that level once the higher end players are moved up and out of their rating.

The larger numbers in 2.5 & 3.0 of bump ups are expected too. Players that are "beginners" will improve at a more rapid rate than the "veterans" that play in higher ranks.

I believe that last year was a sweeping movement by USTA to try to correct years of lack of moving enough players. Without having studied data from years past, I bet that your area had 2-3% pre-2009 movement and that 5-6% is better in line with the actual ability of players.

ian2
11-30-2010, 11:48 AM
Those numbers look pretty normal and expected to me. Seeing as players generally increase in ability over time & rarely decrease in ability it shouldn't be expected to see much in terms of downward movement.

I agree. I wasn't saying there should have been more downward movement. I just stated that there was very little of it. As expected.

The larger numbers in 2.5 & 3.0 of bump ups are expected too. Players that are "beginners" will improve at a more rapid rate than the "veterans" that play in higher ranks.

I agree that beginners improve at a faster rate. However please remember that DNTRP is generated not based on "improvement" but based on the match results against other players who are also improving. I looked up a number of 2.5 and 3.0 players who were moved to 3.0 and 3.5 respectively. It takes less than .500 record (and I mean not just win/loss ratio but the actual scores) to be bumped up from 2.3 to 3.0. This is obviously evidenced by the stats: 60% of 2.5 were moved up. Same trend at 3.0 though not as drastic. Basically it takes much less in terms of the results to be moved up from 3.0 to 3.5 than to be moved up from 3.5 to 4.0. This is not my opinion, it's obvious statistically and proved by looking at individual records. "The larger numbers in 2.5 & 3.0 of bump ups" are NOT expected; they result from USTA manipulating the algorithm selectively for different levels this year.

I believe that last year was a sweeping movement by USTA to try to correct years of lack of moving enough players. Without having studied data from years past, I bet that your area had 2-3% pre-2009 movement and that 5-6% is better in line with the actual ability of players.
It's impossible to tell what the rate of movement was in, say, 2008. At least I don't see how it could be determined using the tennislink drill-down interface. Maybe someone published such stats at this forum back at the time... I'll try to search. You might be right about 2-3% pre-2009 movement. However my feeling (unsubstantiated) is that the movement at 3.5+ levels used to be higher than 6% prior to 2009.

polski
11-30-2010, 01:37 PM
I agree that beginners improve at a faster rate. However please remember that DNTRP is generated not based on "improvement" but based on the match results against other players who are also improving. I looked up a number of 2.5 and 3.0 players who were moved to 3.0 and 3.5 respectively. It takes less than .500 record (and I mean not just win/loss ratio but the actual scores) to be bumped up from 2.3 to 3.0. This is obviously evidenced by the stats: 60% of 2.5 were moved up. Same trend at 3.0 though not as drastic. Basically it takes much less in terms of the results to be moved up from 3.0 to 3.5 than to be moved up from 3.5 to 4.0. This is not my opinion, it's obvious statistically and proved by looking at individual records. "The larger numbers in 2.5 & 3.0 of bump ups" are NOT expected; they result from USTA manipulating the algorithm selectively for different levels this year.


My point is that, beginners generally enter the NTRP system at a 2.5 or 3.0 level. They probably need to be moved up quicker, different algorithm or what not, so that the lower end doesn't become clogged. If they were to all be moved a the same rate, it would be miserable for people new to the game to try to enter against vastly more experienced players.

I assume that the majority of your 2.5 & 3.0 players will be fine at their new level. I never have met a 2.5 player that can't compete at a 3.0 level after 2 years of competing at 2.5. Same goes for 3.0's moving to 3.5...frankly, most embrace it as a sign of respect.

And no, I don't work for USTA. I just value fair competition for all.

J_R_B
11-30-2010, 02:01 PM
I didn't run any numbers, but it seems like there were a lot of bumps from 4.0-4.5 around here. Last year, they skimmed all of the best players that really should have been moved up (plus just a couple who probably weren't really 4.5 players). This year, there were fewer players that deserved to be bumped but a lot got bumped. Do the math - there were far more people bumped this year that really don't belong at 4.5 and will get killed all year.

ian2
11-30-2010, 02:20 PM
My point is that, beginners generally enter the NTRP system at a 2.5 or 3.0 level. They probably need to be moved up quicker, different algorithm or what not, so that the lower end doesn't become clogged. If they were to all be moved a the same rate, it would be miserable for people new to the game to try to enter against vastly more experienced players.

I assume that the majority of your 2.5 & 3.0 players will be fine at their new level. I never have met a 2.5 player that can't compete at a 3.0 level after 2 years of competing at 2.5. Same goes for 3.0's moving to 3.5...frankly, most embrace it as a sign of respect.

And no, I don't work for USTA. I just value fair competition for all.

Good points... once again, I agree with you. I even believe (gasp!) that whatever "tweaks" USTA attempted in the last two years were done for the good of the game in the long term. Still, for those who see NTRP rating as a measure of progress, the inconsistencies in how the algorithm is applied year-to-year and level-to-level combined with intentional lack of transparency are disappointing. So are some demonstrably absurd outcomes resulting from the deficiencies in both the algorithm itself and its application.

dizzlmcwizzl
11-30-2010, 04:49 PM
Delaware district had almost 10% of players moved from 4.0 to 4.5

Although the top three teams were decimated. We lost a combined 21 of 68 players on the top three teams. The rest of the teams in the league lost at most 1 player each.