Monday, February 20, 2017

What about by NTRP level? More interesting league stats on bumps for players that go to Nationals

I've written about some general stats and stats by gender for the bump rates of players on teams that go to Nationals, now it is time to break it down by level.

And this is where it starts to get more interesting as players improve at different rates at different levels, in fact you'd expect the lower levels to have more bumps up as there is more room and a better chance players will improve quickly.  As you can see below, that is generally, true, but there are a few surprises.

As before, we'll start with 18+ and look at players on Nationals team rosters.


Here we do see that the level with the most bumps up is the 2.5 level with nearly half of the players on Nationals rosters being bumped up.  But somewhat surprisingly, there is a higher percentage of 3.5s bumped up (43.1%) than 3.0s (38.7%).  After that, it does drop as you might expect, but still 18% of 4.5s and 15.8% of 5.0s get were bumped up.

Here are the same stats for 18+ for those that played in matches at Nationals.


Looking at those that actually played at Nationals we see the ~5% increase across the board, but the 3.0s still have a smaller percentage bumped up than the 3.5s.

Moving on to 40+.  First the rostered players.


Here we don't see a clear trend that lower levels end up having more bumps up with 3.0 thru 4.0 all being right at 33-34%.  There is a drop at the 4.5 level though.  This does make sense in a way as you'd expect the 40+ crowd to be more likely to be at their right level, even at the lower levels, and it is more consistent in the number of players that improve regardless of level.  That is until you get to 4.5 where there are fewer players capable of getting to 5.0.

Then for completeness, the 40+ players that played at Nationals.


We again see the ~5% increase, although perhaps surprisingly the 4.0 level jumps all the way to 40%.

I know, your next request is to see by level and gender.  That is coming!

Sunday, February 19, 2017

More analysis on being bumped if you go to Nationals - Men vs Women - Interesting Tennis League Stats

I wrote yesterday about the bump percentages for players that are on teams that go to and play at Nationals.  I only looked at the high level stats and naturally it can be broken out a variety of different ways.

The first is by gender as you wouldn't necessarily expect men and women to be bumped at the same rates.

Here are the same charts as the previous analysis, but split out by women and men.  First the 18+ on Nationals rosters.


Here we see that the women have a far larger number of players rostered on Nationals teams, but the percentage that are bumped up is quite a bit lower at 28.6% vs the 34.7% for the men.  Either the men are more likely to be improving, or the men in 18+ have more of a tendency to sandbag.

Then those that played at Nationals.


The women still have more players that played at Nationals, but the gap in what percentage  are bumped shrinks a bit as the women go up to 35.4% and the men nearly get to 40%.

Moving on to 40+.


Women's rosters are still bigger, but not the same way they are in 18+.  And here the women are the ones bumped up at the far higher rate, 32.5% while the men are barely over a quarter at 26.1%.  This may be stereotyping, but it appears women in the 40+ crowd are more likely to still be improving while the men are more likely to have plateaued.  Or maybe over 40 women are more likely to be sandbaggers?

And looking at those that played at 40+.


Both bump up percentages go up significantly, the women to 37.6% and the men to almost 30% at 29.7%.

What do you think?  What observations do you have?

Saturday, February 18, 2017

If you go to USTA League Nationals, will you be bumped up? Interesting Tennis League Stats

A fairly commonly held belief is that if a player goes to Nationals, they will, or should, be bumped up.  We know this doesn't happen though, and it causes much consternation among a few that think it is unfair that someone goes to Nationals and doesn't get bumped up and so has a chance to go back again at the same level.  There are of course other rules that preclude the same team going back at the same level, e.g. the move-up/split-up rule where only three members from a team that goes to Nationals can be on the same team at the same level the following year.

But back on topic, I did some research to see how many Nationals players actually do get bumped up so we could see some real statistics on the subject.

First, I took a look at all the players on the roster of an 18+ Nationals team that held a 2015 year-end C rating and ended 2016 with a C rating.


This revealed that 67.3% stayed the same level, 31.2% were bumped up, and perhaps surprisingly, 1.5% were bumped down!

Now, some teams have large rosters and not everyone is at the top of their level and "Nationals caliber", and not all players will actually go and play in a match.  Here is how it looks if just those that played in a match are included.


Here the bump up rate goes up to 37.2% while the stay rate is 61.9% and just 0.1% of these are bumped down.  This probably makes sense, the players on a roster that don't play are likely lower rated so removing them the rate goes up, but the bump up percentage is still below 40% which goes against the perception that if you go to Nationals you will be bumped up.

Second, I did the same for 40+.  Here is the rostered player chart.


For those on Nationals rosters, 67.8% stayed the same, 29.5% were bumped up, and 2.7% were bumped down.

Then the players that played at Nationals.


Of those that played, 63.9% stayed the same, 33.9% were bumped up, and 2.1% were bumped down.

These 40+ numbers probably make sense, the bump up rate is lower in the first place as those 40+ are less likely to be improving as much and the bump down rate is higher as there are older players more likely to be on the decline.  And we see a similar small increase in the bump ups looking at the players that actually play.

Note, that this analysis did not include self-rates, and the self-rates that make it to Nationals could very well have higher bump up rates.  I'll try to do this analysis soon.

Also, even just the 18+ and 40+ analysis above can be sliced a variety of different way including by gender, by level, and also looking at the teams that made the semis.  If there is interest, leave a comment here or on Facebook and let me know what you'd like to see!