Wednesday, April 27, 2016

How many matches do captains play? - Interesting Tennis League Stats

I recently wrote about what court captains play on the most, so I decided to look at how many matches captains play in.

The chart below shows how many captains played various percentages of their teams matches.


The percent range with the largest count is the 60-79%, so a lot of captains play in 2/3 to 3/4 of their team's matches.  Remarkably, there are some that play in none, and some that play in less than 20% of their team's matches.  And there are perhaps a surprising number that play in every team match.

If we look at the counts by range for non-captains we see it is a bit different.


Here the range with the most is 40-59%, so a bit lower.  And there are more than play in none relatively speaking, but fewer that play in all of the matches.

So it is safe to say that captains do play themselves more on average than they play the non-captains on the team.  Captaining comes with benefits! :)  This is probably no surprise, but the charts certainly confirm it.

As a captain, are you likely to play yourself more?  Or if you aren't a captain, does your's play themselves more or less than the rest of their roster?


Sunday, April 24, 2016

What court do captains play on most often? - Interesting Tennis League Stats

While watching a USTA League match this evening, I got to wondering where captains happen to play themselves most often.  I've written in the past about where captains rate compared to the average by gender and level, so I thought it would be interesting to look at this.

Just looking at 18+ and 40+ matches from 2015 where it was a 2 singles and 3 doubles format, it breaks out as shown below.


For both men and women, captains played most often on court 1 doubles and way more often on a doubles court than singles.  What is a little interesting is that women captains appear to play on the top courts at a higher rate than men; women play court 1 or 2 doubles 39% and 29% more often than 3 doubles, while men play court 1 or 2 doubles 14% and 16% more often than 3 doubles.

From the earlier analysis on strength of captains, they were not generally stronger than average other than at the 3.0 level, so it appears captains either feel they are stronger than they are and play themselves on what is perceived to be the stronger courts, or they are more willing to take one for the team and let other players get wins on court 3.

If you are a captain, what court do you typically play on and why?

Friday, April 22, 2016

USTA PNW Northwest Washington (Seattle) releases 2017 schedule

We are just into the heart of our 18+ and 40+ season for 2016, and the 2017 schedule has already been published for the Northwest Washington district of the Pacific Northwest section of the USTA.

You can see the PDF on the NWW section of the PNW site, but here is the summary of what has changed or is new.

The Adult 55+ and Mixed 18+ leagues continue to be early start and begin in early September with playoffs at the beginning of December.

In the past, including this year, Adult 18+ started right away in January and then 40+ started 1-2 months later and they leagues overlapped for much of the season with both local playoffs in June.  This is all changing.

The Adult 40+ will start right away in early January instead of 18+ like it has been in the past few years, and 40+ local playoffs will be in April rather than in June.  This then allows Adult 18+ to start in late March and have local playoffs in June with very little overlap with 40+.

The One Doubles schedule is much the same as it is this year as are the Mixed 40+ and 55+.

I think these changes make sense as overlapping leagues and both playoffs being in June with vacations and graduations was tricky.

What do you think?  Do you like the changes?

Tuesday, April 19, 2016

Playing in multiple areas, districts, and sections - Interesting Tennis League Stats

In generating Estimated Dynamic NTRP Rating Reports, I often see players that play in more than one are, some in more than one district or state, and even some that play in multiple sections.  Given the geography of the United States and how the section/district/area boundaries are defined, it can happen more often than you think, although it is typically in more densely populated areas.

For example, in the Mid-Atlantic section, Maryland, Virginia, and Washington D.C. are each separate districts and are close enough that some players will play in all three.  And in Maryland each county is a separate area so players can easily play in several.

Or you have sections like Middle States and Eastern, or Eastern and New England, that are very close and players can easily play in multiple.

Even in a less densely populated state like Washington, we have separate districts for Northwest and Southwest Washington, and some players will play in both.

To get an idea of how often it happens, I took a look at the 2015 league year 18+, 40+, and 55+ leagues to see what some of the largest numbers were and if there were any surprises.

First, looking at sections, the vast majority of players, over 99%, play in only one section.  But over 2,300 played in two sections in 2015 leagues, and 8 players played in three sections.  Of those 8, most appear to be situations where someone winters in the south, for example, there were several that were Eastern/Middle States/Florida or New England/Eastern/Florida.

Next, looking at districts (states in Southern), the most played in was 4, and that occurred 50 times.  There were nearly 600 that played in 3 districts and over 7,000 played in 2, but the vast majority play in only 1.  I was actually a little surprised that the largest was only 4, but districts are pretty big areas, so perhaps more would be unexpected.

Last, looking at areas, there were two players that played in a whopping 7 and nine that played in 6.  Both of the players that played in 7 did so in Mid-Atlantic, both predominantly using Maryland's separate counties to do so, but did both include NOVA.

How many different area/districts/sections do you play in?


Friday, April 8, 2016

What is the best roster size for USTA League? Interesting Tennis League Stats

Being a captain for a USTA League team is not an easy job.  One of the first things they have to do is assemble a roster, and a common question from new and even experienced captains is how large a roster they need to have.

On one hand, you want to have a large roster to make sure you have enough players and never have to default a court.  And depending on the time of year, work schedules, vacations, kids, etc. this can be tricky.

On the other hand, you want a small enough roster so that every player can get a reasonable number of matches.  If you happen to have a schedule with a lot of matches, that eases this burden, but some leagues will only have 8 or 9 matches on a schedule which isn't a lot.

Then the captain has to balance trying to field a strong team and playing the best players vs getting the weaker players ,who are needed to fill out the roster, enough matches to make it worth their while to be on the team.

Some basic math tells us that for a standard 5 court match (2 singles and 3 doubles) there are 8 players required, so an 8 match season would require 64 player spots.  A roster of 16 would allow each player to play 4 matches if evenly distributed.  If a team has a 12 match schedule, that is 96 player spots and then the same 16 player roster would allow 6 matches per person.  But if a team with an 8 match season had 20 players, that is barely 3 matches per player.

So what is the right size?  It really depends on a lot of variables noted above, but we can look back at 2015 to see what sort of statistics we can find on typical and exceptional roster sizes.

What I did was to look at the 2015 18 & over league to look at both the size of rosters but also the number of players that actually got in a match during the year.  Here is what I found.

First, here is a chart showing the number of teams at each roster/played count.


So it appears the sweet spot for roster sizes is 15, that is the most common size, but the most common number of players that play in a match is 14.  The average roster size was 15.0 while the average number of players that play is 14.5.  But you can see it is not uncommon to have rosters in the high teens and even a fair number of teams have rosters in the low 20s.

Looking at some averages and extremes, the smallest roster was 8 players and there was a team that only had 6 play, so they were always defaulting a court.

On the other end of the spectrum, there was a team with a roster of 38!  There was also a team with 33, a couple with 32 and 31 each, and the frequency becomes greater as you continue to get smaller.  The 38 player team somehow had 36 players play as well!  The next highest number of players playing was 29.

This 38 player team did have 12 matches, so that helped, but the majority of players played just 2 or 3 matches.  Four players played 4 and just one player played in 5.

What do you think the right roster size is for a normal 2 singles and 3 doubles league?

Thursday, April 7, 2016

The Pacific Northwest Section releases new rules for 2017 Early Start Leagues

I wrote a few weeks ago about early start ratings going away for 2017 leagues and the implications of that, and the PNW section just sent an e-mail out with more details, specifically this link has the new rules.

But here is the summary:

  • When registering for 2017 early start leagues, players will use their current NTRP level (2015 or earlier valid year-end level) as there will not be any early start ratings published.
  • If a player is bumped up at 2016 year-end, they may continue playing at their 2015 NTRP level on the 2017 early start team subject to:
    • Any player with a year-end rating that is greater than the clearly above level (strike threshold) will not be eligible to continue playing at the lower level but matches played will stand.
    • Any players that are bumped up at 2016 year-end will not be eligible to play at Nationals unless they were playing at the 2.5 level and their rating did not meet the clearly above level at 2016 year-end.
    • Each section will determine if players that are bumped up at 2016 year-end will remain eligible to play at the lower level for championships played after year-end ratings are published.

The above is all consistent with what I wrote earlier, the new item being the extra clause for the 2.5 level.  And it appears the PNW has not yet decided what the rule will be for playoffs/championships played after 2016 year-end ratings are published, but will do so prior to sign-ups for early start leagues.

What do you think the rule should be for championships/playoffs?  Should players be ineligible immediately on bump up since they can't go to Nationals anyway?  Or should a team be allowed to go as far as possible with their full roster?

As I wrote earlier, my vote would be to not allow players to continue at their old level as soon as reasonably possible, since that is fairer to other teams and will allow the section to be represented by a team that can field the same/similar team at Nationals as won the section.

Sunday, April 3, 2016

Do stronger players in USTA League play singles or doubles? Interesting Tennis League Stats

In my work generating reports but also in questions I get in e-mail in response to my NTRP Ratings FAQ, a question I often get asked is if a player should play singles or doubles to improve their rating the most.  There is no clear answer, it all depends on the ratings of the players in the match and the score, and some players are simply better singles players than doubles or vice versa, so there is no golden rule.

However, I've told lots of people that singles can offer more opportunity to improve your rating as stronger players typically play singles.  This is the conventional wisdom, that the best players play singles, so beating them will result in a better match rating than beating lower rated players on a doubles court.  There are exceptions to this of course as the rating of your partner matters too, but I wanted to test this conventional wisdom to see if it was indeed true.

So, I looked at 2015 leagues in the 18 & over division, and used my Estimated Dynamic NTRP Ratings as of the end of 2015 to calculate the average rating by court for each gender.

In the charts below, I show the average for each NTRP level showing 1S, 2S, 1D, 2D, then 3D in the groupings.  I've also normalized the ratings to be relative to the bottom of each level so that the relative strength within the levels can easily be compared.

First the men.

From the chart, it is pretty clear that the strongest court on average is court 1 singles, and that the lower courts (2 and 3) have lower rated players on average than higher courts.  What is interesting is that court 2 singles is on average lower than court 1 doubles, but it is stronger than court 2 doubles.

Another interesting datapoint is that at the 5.0 level, there is a little aberration from the norm and court 3 doubles is actually slightly stronger than court 2 singles or doubles.  Note that a lot of 5.0 leagues are just 1 singles court and 2 doubles courts though so that likely skews these numbers some.

But back to singles vs doubles, this chart would seem to largely support the notion that stronger players play singles, except that court 1 doubles is usually stronger than court 2 singles.

Another interesting observation is that as the NTRP level increases, the average rating for every court is less and less above the bottom of the level.  This sort of makes sense, since I used year-end ratings, these reflect how much players have improved during the year, and 2.5s and 3.0s are going to improve a lot more than 4.5s and 5.0s.  Also, at the higher levels the distribution of players is going to be at the lower end of the level.

Last, it appears captains do generally play straight-up and play stronger players on the lower numbered courts.  The USTA does not require this, captains can and do stack their courts to try and steal a win, but it doesn't look like stacking is done nearly enough to skew the data much at all.

Now the women.

The women look largely the same, but sometimes court 2 doubles has a higher average than court 1 singles which was not the case for the men.  The women also seem to fall off a bit more on court 3 doubles than the men.

So there is a slight bias towards singles being stronger than doubles, but not nearly as distinct as for the men.

Also, the women seem to have generally slightly lower averages across the board on each court.  My hypothesis would be (from experience doing reports and looking at where players play) that more women play up than men and so this pulls the averages down.  But I'll do some research to confirm this soon.