Tuesday, October 24, 2017

What if it hadn't rained in Mobile for 2017 USTA League 40+ 4.5+ and 18+ 2.5 Nationals? A solution to picking a champion, EDNR

It rained in Mobile, AL this past weekend where USTA League Nationals were being held.  While the tournament officials were creative and rescheduled the 4.5+ semis to late Saturday in the hopes they could get the final in Sunday morning, the rains came early and the finals were not played/completed.  Worse, for the 18+ 2.5 women, because teams had already played twice on Saturday, the semis were not rescheduled resulting in that event not even determining the finalists.

How do they declare a champion then?  They don't, and punt and declare everyone co-champions.  Meaning, rather than three teams being champions, there were eight!  All four 2.5 women semi-finalists and the four 4.5+ finalists for the men and women.

That would be very unfulfilling, and from what I hear, folks left disappointed.  I can imagine not getting to finish things on the court would leave a sour taste in your mouth, but I heard that teams still had fun.

Let us ignore for a minute that playing Nationals in Mobile, one of the wettest cities in the country, is a questionable move on the USTA's part, and also that Nationals are also being played in Florida this year, another not exactly dry location where rain caused delays and altered playing formats a few weeks ago, and instead just look at how a single champion could have been determined when it can't be decided on the court.

A first thought it to use some sort of tie-breaker like is used in determining flight champions.  The challenge is, flight champions play each other and the same opponents, so courts won/lost and head-to-head make sense.

The USTA does already have an existing way to compare teams between flights though as they have to do it when Nationals has three flights and a wildcard makes the semis.  In this case, my understanding is that they use percentage of games won.  This can work when a flight may have a different number of teams and is finer grained than looking at courts won/lost or sets won/lost, but it still suffers from a team in a weaker flight having an advantage over the second place teams in the other flights.

Since I calculate Estimated Dynamic Ratings, my mind naturally starts to think about how ratings could be used.

One way of using them would be to look at the ratings players achieved during Nationals play, the idea being that those that the team that had the highest average match rating had shown themselves to be the best team.  This is similar to percentage of games won in that it is result based from Nationals, but it factors in the strength of the opponent as well as the score instead of looking just at the score.

If we did this with what happened in Mobile, the champions would have been:
  • 2.5 women - Pacific Northwest had the highest average at 2.80 vs Florida's 2.76, Southern's 2.70, and Missouri Valley's 2.60
  • 4.5+ women - Southern would have a very narrow win over Florida 4.46 to 4.45
  • 4.5+ men - Middle States would beat Northern California 4.62 to 4.52, but interestingly Florida had the highest average match rating at 4.63 but lost to NorCal 3-2 in their flight

But using an average rating achieved for a team isn't terribly satisfying, that rewards a deep team and ignores the match-ups a captain may have been able to get via line-ups to "steal" a 3-2 win.

As you might expect, I have a better solution!  Now, keep in mind I'm a bit of a propeller head so this appeals to be, but I think it makes sense and would be equitable and objective.

Ideally the captains fill out their line-ups and the match is played on the court.  If the match can't be played on the court though, the captains can still fill out the line-ups as they would have and then the USTA could use the currently calculated dynamic ratings through Nationals play to play the match "on paper" giving the court win to the player/pair with the higher rating(s).  This would even work with the 2.5 women where they had to play the semis as the captains could fill out their line-up for the semis and see who "wins" and then the winners could do it again for the final.

Now, the USTA doesn't tell us the dynamic ratings for the players, but lucky for us I calculate estimated ratings so we can use those!  Better yet, I actually have a match simulator to compare line-ups and identify who the winner would be!

So what I'll be doing for the levels not completed in Mobile, and any other Nationals that aren't able to complete their finals, is to identify who the Estimated Dynamic NTRP Rating (EDNR) National Champion is.

I'll be doing these in subsequent posts and will use my best guess at line-ups unless there is better information available.  If any captain of one of the teams wishes to contact me and let me know what their line-up was or would have been, please do.  In lieu of this, for the 4.5+ men and women I will use their line-up from their semi-final match.  Otherwise I'll use what appears to be the most common or likely line-up.

Stay tuned, but feel free to let me know what you think of this approach.

No comments:

Post a Comment