Sunday, December 27, 2015

2015 USTA League team winning scenario percentages - Interesting tennis league stats

A question was asked on Talk Tennis about how often a team wins a match by winning the three doubles courts while losing the two singles courts.  I calculated the answer to that (5.9%), but that naturally caused me to look at all the other scenarios a team could win a match with.

For this analysis, I looked at all 2015 championship year 18 & over and 40 & over matches that were a 2 singles and 3 doubles.

To start, there are three scores a team can win a 2 singles / 3 doubles match with: 5-0, 4-1, and 3-2.   There is only one way to win 5-0, five ways to win 4-1, and ten ways to win 3-2.

Here is a chart showing the percentage each one occurred along with what would be expected based on the number of permutations for each score.


Somewhat surprisingly, teams winning 5-0 occurs a lot more often than expected.  I've done double checking of the data and I am not including defaults or retirements, so I think it is accurate.  Even 4-1 occurs more often than expected while a 3-2 score occurs most often, but less than expected.

But what about each specific winning scenario?  There are 16 of these, and the chart below shows how often each one occurs.


The x-axis is showing the courts won in the order singles 1 and 2, then doubles 1 thru 3, e.g. 11110 is winning all courts but 3rd court doubles.

Since there are 16 potential results, if each one was equally likely, you'd expect each to be at 6.25%, but we see winning all the courts is again a lot higher, and the 5 4-1 scenarios are all over, and then all of the 3-2 scenarios are under.

I'm not sure exactly what this tells us other than close 3-2 matches are less common than you'd expect.

What do you think?  Is this consistent with what you see in your league?


Sunday, December 20, 2015

2013 - 2015 USTA League participation by year, division/league, and gender, USTA League players are aging - Interesting tennis league stats

I took a look at how participation in USTA League has changed the past three years, specifically looking at the number of unique players playing in all of the advancing 18 & over, 40 & over, 55 & over, and 65 & over divisions/leagues.  The natural next step was to look at the numbers for each division, so here goes.

First, here is the 18 & over division.


The trend here is similar to the overall trend, the drop here being just over 3% since 2013.

Here is the 40 & over division.


Interestingly, while the totals for this division are lower than the 18 & over as you might expect, the unique players in this league has grown each year!  Participation is up nearly 7% since 2013.

An explanation for this might be that as some of the 38 and 39 year olds have turned 40, they've elected to only play 40 & over and there have not been enough new under 40 year-olds joining to fill the void.  And those that turn 55 seem to continue to play 40 & over.  Also, the 40 & over division was new for 2013 so perhaps it took some a year or two to get on board with it.

Let us see if the trend continues in 55 & over.

Sure enough, the trend continues.  This group is up over 4% since 2013.  As the 53 and 54 year olds have turned 55, they keep playing at a higher rate than they get too old or injured and stop it would appear.

Given all this, it would be interesting to see what the average age of these players is.  Alas, I don't have that data, but I would hypothesize that the average age is increasing which would indicate that existing players keep playing as long as they can, but the new younger players are not joining and playing USTA League enough to replace the older players that do end up not playing at some point.

What do you think?

2013 - 2015 USTA League participation by year and gender - Interesting tennis league stats

I see comments periodically saying that participation in USTA League has been declining the past few years.  Naturally, that gave me reason to take a look.

Using the data I have from calculating Estimated Dynamic NTRP Rating Reports, I took a look at how many players have played in the 18 & over, 40 & over, 55 & over, and 65 & over Adult divisions regular season.  These number do not include players that only played Mixed, or only played in Tri-Level, Combo, Singles, or other non-advancing leagues/divisions.


First, kudos to the USTA for organizing leagues that get nearly 300K players participating.  I'm sure this is a small fraction of all the tennis players playing in the United States, but is still a good sized number.

But to answer the question about participation, there does appear to be a small but consistent decline the past few years.  It may be hard to see in the chart, but total participation was down about 4,000 from 2013 to 2014, and was down another 2,000 this year.  So just over 2% from 2013 to 2015.

And the decline is pretty consistent across genders, although men's participation almost stayed flat from 2014 to 2015.

With all data and analysis, we need to understand what we are looking at and this is looking just at participation in the advancing leagues as noted above.  It could very well be that while these have declined, the other non-advancing leagues haven't and could even have grown.

As to why there might be a decline, I can only speculate.  There could just be fewer people playing tennis, or perhaps league fees have gone up and driven a few away, or some have decided they don't want/need to play in the advancing leagues and taken their play to other leagues/divisions.

What do you think?

Saturday, December 19, 2015

Christmas Estimated Dynamic NTRP Rating Reports are back! The perfect last minute gift for the tennis nut

Once again, I am generating special PDF versions of Estimated Dynamic NTRP Rating Reports, perfect as a stocking stuffer or last minute gift for that hard to buy for USTA playing tennis friend or family member.  The PDF of the report will be provided at no extra charge and you'll still get my explanation of the report in the e-mail I send with the PDF.

Here is what the PDF will look like.


Contact me to request a report or with any questions.  And Happy Holidays!

Friday, December 18, 2015

Tracking the activity of your dog doing agility - FitBark dog activity monitor

I recently wrote about tracking my own activity when playing tennis using a Misfit Flash activity monitor.  But what precipitated it all was getting an activity monitor for our dog and that piquing my interest in tracking my own activity.

My wife and I have a very energetic Cocker Spaniel, and realized pretty early on that we needed to find something for him to do.  My wife ended up getting into agility with him, first for fun, but now competing in agility trials, and they are both doing great.

A natural question we had was how active our dog really was.  We came across an ad for pet activity monitors, and after doing some research purchased a FitBark Dog Activity Monitor.

The small bone shaped device attaches to your dog's collar and monitors his activity just like an activity monitor you would wear does.  What makes it really useful and interesting is the app you install on your smart phone that "fetches" the data from the device and displays it for you.

The data that is retrieved is turned into points, and a daily point goal is set based on the dog's breed, age, gender, weight, etc. (but it can be overridden).  In the image above, the details for the day are displayed and then a summary of the daily totals for the week.  The green line is the daily goal, which you can see the our dog is exceeding his daily goal on a regular basis and on this day was over 50% over it.

But it isn't just the total for the day, but the hourly view you get in the chart and the break-down of how much time was spent resting, active, and playing.  The hourly view is great as we can see exactly how much activity he gets when we go for a walk or go to the park to chase the ball, or in our case how many points he gets from a run through an agility course.

The app has other features to track a "pack" of dogs, compare your dog's activity with other dogs of the same or similar breeds/sizes, share the information with friends, family, or your vet, and even link it with your own activity monitor should you go running or biking with your dog.  The app has even been updated several times in the couple of months we've been using it so they continue to enhance and improve it.

And the FitBark even collects data when you aren't home, so you can see exactly how active your dog is when no one is home.  There is even a WiFi station you can get that fetches the data while you aren't there and uploads it to the cloud so you can run the app and see the up to the minute activity and points.

We have been very pleased with the FitBark and would highly recommend it to any agility dog owner, or any dog or pet owner, as it provides new insights into the activity and health of your pet.

Do you have a FitBark or other activity monitor for your dog?  I'd love to hear your comments.

Tracking your activity when playing tennis - My experience with a Misfit Flash

This post is going to deviate a bit from discussing NTRP ratings and USTA League stats, but is still about data related to tennis.

Personal activity monitors, e.g. Fitbit, Garmin Vivo, Nike Fuel, etc. have become very popular recently.  I succumbed and got one, a Misfit Flash, in part because I'd gotten a FitBark activity monitor for my dog (see more here) and also being a data junkie, getting data about my activity is a bit intriguing.

I have not used many other activity monitors, and this is not intended to be a full review of the Flash, but it is a fairly inexpensive device does provide the basics of monitoring your steps, calories, and miles.  It also accumulates points for you and measures your progress towards a daily goal.

The points it calculates are based on a variety of factors, one being the type of activity you are doing.  You are able to track a specific activity and identify what the activity type is.  You also tell it where you are wearing the device (wrist, shoe, waist), and when you do this, it calculates your points accordingly.  For example, the same activity it monitors for a period will result in more points if you tell it you were running vs walking.

But what I really care about is what it says about playing tennis, and thankfully, there is a specific activity type for tennis.  So I mark the start and end of when I play, tell it the activity was tennis, and voila, it tells me how many points I accumulated for the activity.

The image above shows the total for a day in which I played twice for a total of about 3 hours.  There was about 1.5 hours of singles and 1.5 hours of doubles.

The image to the left shows the specific activity when I played singles on this day and you see I garnered a bunch of the points from it.

I do notice that I get more points from singles than doubles.  I probably get about 900 points per hour from doubles and 1,200-1,400 per hour from singles which all makes sense.

All in all, I'm very pleased with my Misfit Flash.  I'm sure other trackers do more and have more features, and the Flash has some quirks, for example there is no display, just a dial with LED's and a single button.  But you can monitor your daily progress, start/stop an activity, and see the time with something as simple as this.

If you use an activity tracker when playing tennis, leave a comment and let me know what your experience has been.


Sunday, December 6, 2015

2015 USTA League bump percentages by level, 5.0 men and women - Interesting tennis league stats

This post continues the thread of analyzing the 2015 USTA League year-end bumps up and down.  I've already looked at bump up/down percentages by section and overall, as well as for the 3.0 men and women3.5 women and men4.0 men and women, and 4.5 women and men.  Now I take a look at the 5.0 men and women.

Here are the men.

It is no surprise there aren't many bump ups, but the number of bump downs is a bit surprising, particularly in Hawaii and New England where about half and one third were bumped down!  And a bunch of other sections had more than 15% of 5.0s bumped down.

Here are the women.

The women are similar, but Caribbean and Northern have nearly every 5.0 bumped down, but there are also very small numbers here with hardly any 5.0s in the first place.

So it seems there is definitely going to be a bunch of strong 4.5s in 2016 with all of the 5.0 bump downs.

Saturday, December 5, 2015

2015 USTA League bump percentages by level, 4.5 men and women - Interesting tennis league stats

This post continues the thread of analyzing the 2015 USTA League year-end bumps up and down.  I've already looked at bump up/down percentages by section and overall, as well as for the 3.0 men and women3.5 women and men, and 4.0 men and women, now I move on to the 4.5 women and men.


Like the 4.0 women, there are a remarkably high number of bump downs to 4.0.



The men even have more bumps down than up!  The 4.0 level is going to grow a bit.

2015 USTA League bump percentages by level, 4.0 men and women - Interesting tennis league stats

This post continues the thread of analyzing the 2015 USTA League year-end bumps up and down.  I've already looked at bump up/down percentages by section and overall, as well as for the 3.0 men and women and 3.5 women and men, now I move on to the 4.0 men and women.


Other than Caribbean again, the men look pretty normal with a few more bumps up than down.


The women is pretty interesting though with nearly every section having more bumps down than up!  Is this perhaps an adjustment in response to bumping too many up last year?  If so, was it a natural adjustment because those players couldn't do well at 4.0 or was there an adjustment the USTA did?

2015 USTA League bump percentages by level, 3.5 men and women - Interesting tennis league stats

This post continues the thread of analyzing the 2015 USTA League year-end bumps up and down.  I've already looked at bump up/down percentages by section and overall, as well as for the 3.0 men and women, now I move on to the 3.5 women and men.

As you might expect, things begin to become more balanced at 3.5, but there are still more bumps up than down.

The women have 3-5% of 3.5s being bumped down to 3.0 and 5-9% being bumped up to 4.0.


The men are similar to the women although the Caribbean really has a high percentage and makes the chart look different.

Friday, December 4, 2015

2015 USTA League bump percentages by level, 3.0 men and women - Interesting tennis league stats

This post continues the thread of analyzing the 2015 USTA League year-end bumps up and down.  I've already looked at bump up/down percentages by section and overall, now it is time to drill in and look at the same data by section and level.  To start, I'll look at the 3.0 men and women.

As you might expect, at the 3.0 level there are a lot more bumps up than down as not many players are bumped down to 2.5 from 3.0.


 Several sections had around/over 20% of their male 3.0s bumped up to 3.5, and every section had over 10% of them.  Caribbean had over 35% of their 3.0 men bumped up.


For the women, there are actually a few more players bumped down, and fewer bumped up with the top sections around 10-15%.

Are captains in USTA League typically stronger or weaker players? - Interesting tennis league stats

There was recently a question on Talk Tennis asking if captains are generally good players.  Captains are an important part of USTA League and often don't get credit for what they do, and not only handle captaining but play as well.  It seemed interesting to research if they tended to be stronger/weaker for their level, so I did.

I looked at my Estimated Dynamic NTRP Ratings for 2015 and for the 18 & over and 40 & over divisions, for each gender, what the average dynamic rating was for captains vs the average for players rostered by level and gender.  Here is what I found.

For each case, I'm showing the average rating for captains and then the average for players rostered on teams at each level.

Men 18+

3.0 - 2.78 vs 2.74
3.5 - 3.23 vs 3.24
4.0 - 3.68 vs 3.71
4.5 - 4.11 vs 4.15
5.0 - 4.52 vs 4.57

If anything, at levels other than 3.0, the captains are rated slightly lower.

Women 18+

3.0 - 2.76 vs 2.72
3.5 - 3.18 vs 3.18
4.0 - 3.59 vs 3.60
4.5 - 4.01 vs 4.01
5.0 - 4.47 vs 4.48

The gaps here are a lot smaller, other than 3.0 it is in statistical noise.

Women 40+

3.0 - 2.76 vs 2.73
3.5 - 3.19 vs 3.19
4.0 - 3.60 vs 3.61
4.5 - 4.05 vs 4.05

Very similar to the 18+, virtually the same other than 3.0.

Men 40+

3.0 - 2.78 vs 2.78
3.5 - 3.24 vs 3.26
4.0 - 3.68 vs 3.72
4.5 - 4.15 vs 4.21

Similar to 18+, higher levels the captains are a little below the average.

So if you want to look for any generalization, it would be that higher level men's teams are more likely to have captains rated lower than average.  But for women's teams and lower rated men's teams, it is a lot closer and at 3.0, captains may be stronger.

Thoughts?  What is your experience?

Thursday, December 3, 2015

2015 USTA League bump percentages by section - Interesting tennis league stats

The 2015 USTA year-end ratings are out and so analysis can begin.

I'll start with looking at bump up/down percentages by section, first overall, then for male and female.


This looks a lot different than last year (2014) when nearly every section had over 12% of players bumped up and a few were over 20%.  This year, Caribbean is the only section with a bump percentage greater than 10%.

Last year also had most every section with significantly more than twice the bumps up than down, this year the ratio is below 2.0 for most sections.

So this looks a lot more like a normal year and there is not any obvious big adjustments made by the USTA.

We can look at the same chart, but just for the men.

We see Caribbean is even more of the outlier here, nearly every section for the men is below 8% bumps up and ratios to bump down around 1.5.

For the women:


The Pacific Northwest has the highest bump up percentage here, but it is barely over 10%.



Wednesday, December 2, 2015

2015 USTA League year-end bump percentages - Interesting tennis league stats

USTA 2015 year-end NTRP rating levels are out, and now begins assessing what happened this year.

At the highest level, we can look at what percent of players stayed the same, were bumped up, or were bumped down.  Last year, more than twice as many players were bumped up than down (14.3% to 6.4%), but this year, the number of players bumped either way are down and the ratio is a bit smaller with 8.0% being bumped up and 4.7% being bumped down.

Breaking it down by section and state coming soon.