Friday, November 30, 2018

Analyzing 2018 USTA NTRP year-end ratings - Distribution of players by level

Next in the sequence of analyzing the 2018 year-end ratings, we'll take a break of bump up/down percentages and look at how many players there are at each level.

Here are the number of players, men or women, with C ratings at each NTRP level.  I've included 2017 year-end counts as well for comparison.  The percentages on each bar are the percent of the total at that level.


This looks pretty much as you'd expect, the majority of players are at the 3.5 level, followed by 4.0, then 3.0.  It is interesting to see the counts for 2018 year-end slightly lower than 2018 though.  But analyzing declining participation in USTA League play is a topic for another day.

Here is the distribution for the women.


Similar shape, but after the 3.5 level there are more 3.0s than 4.0s, and even more 2.5s than 4.5s.

And for the men.


You can probably do the "subtraction" between the first two charts, but this clearly shows for the men, there are basically the same number of 3.5 and 4.0 men, and very close to the same number of 3.0s and 4.5s.

Stay tuned for more!

Appeal Rules for USTA NTRP Ratings

With 2018 year-end ratings now published, players that aren't happy with where they ended up may take the steps to appeal their rating on TennisLink in the hopes of having it granted.  Some may appeal up so they can play on a team with friends, others may appeal down if they don't think they can compete at the higher level or flights aren't offered in their area at the higher level.

The USTA League regulations list a number of things regarding appeals, but some of the criteria used for appeals is not public.  I wrote a blog five years ago when some of the criteria was included in a communication from the USTA, but I believe the USTA can and does make changes to the criteria, possibly as often as each year, and those rules I wrote about are now dated and no longer valid.  So if you come across that blog, don't bank on it being accurate anymore.

What we do know is what is in the regulations and here is a summary of that.  Note that for the purposes of this article, I'm only addressing the standard auto-appeals.  Manual appeals for medical or other reasons are an entirely separate thing.

First, there are some special rules for players age 60 and over.  Here is the relevant language from the regulations.
2.05E Promotion of Players 60 or Over and 65 or Over 
2.05E(1) Any player who is 60 years of age or older prior to, or during, the calendar year in which such player plays his or her first local league match and has achieved the same rating level or lower for his or her three most recent year-end ratings, without benefit of appeal of the player’s year-end rating, will automatically be granted an appeal (A rating) if promoted, subject to 2.05E(3) below. NTRP Dynamic Disqualification procedures as outlined in 2.04B(3) apply. 
2.05E(2) All players age 65 or over, if promoted, will automatically be granted an appeal (A rating) of their current rating back to their previous valid year-end rating, subject to 2.05E(3) below. NTRP Dynamic Disqualification procedures as outlined in 2.04B(3) apply. 
2.05E(3) Any player who is clearly above level under the applicable Computer Methodology procedures will be denied an appeal of his or her year-end rating under 2.05E(1) and 2.05E(2).

This basically means players over 60 that have had the same C rating for their past three ratings will be automatically granted an appeal down if they are bumped up.

Players over 65 will always be granted an appeal down, except if their rating is too far into the range for the next level.  This latter part (2.05E(3) is new for 2019 and just closes a loophole where a player that really can play at the next level (their rating is above the "clearly above level") could appeal down, now they can't.

Ok, what about the rest of us?  That is covered by a pretty generic item in the regulations.
2.05C Appeal of Year-End Ratings. Year-end NTRP valid computer rating levels may be appealed or denied through TennisLink except for 2.05D Medical Appeals.

That doesn't offer a whole lot, it just says an appeal can be done using TennisLink.

As far as eligibility to appeal, there is this:
2.07 CHAMPIONSHIP PLAYERS. 
A championship player’s computer rating achieved as a result of play in National Championships may not be appealed down following the Championship Year it is received except as in Reg. 2.05D Medical Appeals and Reg. 2.05E Promotion of Players 60 or Over and 65 or Over.

This is just saying that players that go to Nationals are not eligible to appeal down at year-end of the year they go to Nationals, except for those age 60 or over (see above).

What else do we know about eligibility to appeal?  Not much.  In the past there have been documents that mention being within 0.05 of a threshold, or graduated allowances based on the number of matches played, but I have not seen any specifics of what the current criteria are.

If you go online to appeal and it is not granted, you will see a message something like this:
Automated Appeal Decision

Automated appeal requested up for <player name> is denied - Benchmark or out of appeal range.

That simply tells you you weren't eligible or are out of range.  Don't get confused by the "benchmark" terminology, I think that is just left over from when players did get "B" ratings for going to playoffs, but that rating designation is no longer used so you should just interpret it to tell you your appeal is denied with no other special qualifiers.

Note that in all cases, a player that has an appeal granted is subject to 3-strike DQ.  This is more or less the same as a self-rated player in that if your rating gets above the strike threshold three times, you are DQ'd and promoted to the next level immediately.

Should I learn more, I'll add to this blog entry, but I'm also interested in hearing from players that appeal on whether or not the appeal is granted, so don't hesitate to leave a comment or contact me.

Analyzing 2018 USTA NTRP year-end ratings - Bump rates by level and gender

With the 2018 year-end ratings published, I wrote about the general and by section bump rates and also broke that out by gender and section, now to look at bump rates by level.

Here are the bump rates by level for both genders.


As is typical, there are more bump ups at lower levels and more bump downs at higher levels.

Here are how the rates look for the women.


The chart looks remarkably similar.  Let's look at the chart of the men to see if it is any different.



We see there are more bump ups for the men at the lower levels, and at 4.0 the women have more bump downs than up while the menu are still slightly more up.  No big differences though.

More to come.

Thursday, November 29, 2018

Analyzing 2018 USTA NTRP year-end ratings - Overall and by section bump rates by gender

I just wrote about the general and by section bump rates now that 2018 year-end ratings are out, and naturally we want to begin to drill down, so here is how it is broken out by section and gender.

First, nationally, the women had 9.5% bumped up and 4.8% down, while the men had 8.3% bumped up and 4.6% bumped down.  And because Caribbean had a very high bump up rate as we saw in the prior post, removing them can show what happened elsewhere nationally and the women had 9.3% bumped up and 4.8% down, while the men had 7.8% bumped up and 4.7% down.

Now, on to the charts!

Here are the women broken out by section.


We see Caribbean again with a high bump rate and the others pretty consistent.  But removing Caribbean to see more details for the others shows the following.


The additional detail is still pretty consistent, but Mid-Atlantic and Florida both are just over 10% of women bumped up, most others around 9% except for SoCal and Hawaii with significantly lower bump up rates.  All the sections had bump down rates around 4-5%.

Here are the men broken out by section and including Caribbean.


We see Caribbean's 45+% bumped up and all the others.  Here are the others in more detail.


Hawaii and Mid-Atlantic lead the way as the only ones over 9% bumped up, the others are mostly 7-8%.  Interestingly, Hawaii had the highest percentage bumped down too at around 7%.

More to come.

Analyzing 2018 USTA NTRP year-end ratings - Overall and by section bump rates - Caribbean gets rocked!

2018 year-end ratings are out for USTA League players, and they were on-time or even a few hours early based on what I'd predicted.  Thanks to the USTA for getting them out this week as I know in many areas teams are forming now and knowing where players are rated for 2019 is kind of a key ingredient to forming teams.

Now, on to the analysis!

Time allowing, I'll be slicing and dicing the data every which way, but to start we'll just look at some overall and by section bump up/down statistics.

For this analysis, I'm looking at players that had 2017 year-end ratings and got a new 2018 year-end rating.

First, the overall bump rates Nationals across all levels and sections are 9.1% bumped up and 4.7% bumped down.  These are pretty consistent with most years, so nothing out of the ordinary so far.

Next, looking at bump percentages by section, we get the following chart.


We see that Caribbean has hardly any bump downs and over 45% of players bumped up!  Caribbean routinely has a large number of bump ups, but not usually this high.  Perhaps their continued success at Nationals has reaped a larger adjustment this year.  Stay tuned for more detailed analysis, but it will be interesting to see how this affects Caribbean at Nationals next year.

Update: Since I removed Caribbean below, it is also worth noting that with that section removed, the National results were 8.8% bumped up and 4.8% bumped down.

Other than Caribbean, the other sections are all pretty consistent.  Removing Caribbean so we can see a bit more detail though.


We see no section reached a bump up percentage of 10%, Mid-Atlantic was the closest, and Hawaii had the highest bump down rate and one of the lowest bump up rates.

These are pretty consistent from section to section though, so no obvious big adjustments at least at a macro level.  As I dig into gender and level within a section, I expect we'll see more variation.


Wednesday, November 28, 2018

2018 NTRP ratings have been published!

As I write this, it appears the year-end NTRP ratings have been published on TennisLink.  I am seeing players with ratings with a 12/31/2018 date so at least some are now available.

To check what your rating is, you can go to TennisLink and look yourself up, or login and it should show your rating.  Make sure to check the date and that it is 12/31/2018 as that is what all new ratings should show.  If it still says a different date, yours may not be updated yet, or you didn't play enough matches to get a new rating.

Stay tuned for analysis of the ratings, but if anyone has any questions or wants to get a report to understand why they were/weren't bumped up or down, contact me!  And I'm always interested in situations where players successfully appeal, so if you do that and it is granted, drop me a note.

USTA NTRP year-end ratings should be published tomorrow, is everyone excited?

Unless there is some unexpected delay, I anticipate that year-end NTRP ratings to be published tomorrow (Thursday).

Now when tomorrow I don't know.  My guess is late in the day, but who knows, it could be early in the day and first thing in the morning.

In any case, it should be an exciting day as players find out what level they will be at for 2019 so they can begin to form teams.

Stay tuned throughout the next several days as I'll be doing some analysis of the year-end ratings and show bump percentages by section and level so we can look at trends that there may be.

I'll also be making any necessary adjustments to my ratings to correlate with what the USTA did with year-end ratings and then be in a position to start doing reports for 2019.  This includes individual reports, pre-season team reports, or recruiting reports.  Or I can do miscellaneous lists of players meeting whatever criteria you might have.

And of course, I can still do reports for 2018 if you were/weren't bumped and are curious why.

Stay tuned, and try not to check TennisLink every five minutes between now and when ratings are published.

Sunday, November 25, 2018

Cyber Monday Sale - Discounts on all reports

It is Cyber-Monday which means it is time for a sale!

Yes, 2018 year-end ratings should be published later this week, but there are still reasons to get an Estimated Dynamic NTRP Rating Report or any of the other reports I offer.

When year-end ratings are published, you'll know your new (or same) NTRP level, but you won't know anything about how it was arrived at or why it is what it is.  My reports generally predict year-end levels very well and will give you a lot of detail on how your matches rated and why your rating is where it is.

And I can do more than just individual reports, I can do reports for teams or even recruiting reports to help you identify players that meet the criteria you specify to help captains identify players to include on an upcoming team.

Regardless, thru the end of Cyber-Monday, all reports are now 25% off.  If you are interested or have any questions, contact me!

Happy Holidays!

Monday, November 19, 2018

Poll: The conundrum of Early Start Leagues and Ratings

Note: Poll below, please read to the end and participate!

USTA NTRP year-end ratings should be out in less than two weeks and with that, players will know what level they can play at for 2019.

Except some players are already playing "2019" seasons in what are called Early Start Leagues (ESLs).  These are leagues where for a number of reasons (schedule conflicts, court availability, just to get another season in), an area will have a league that is part of the 2019 championship advancement start in 2018.  An example in my area is 55 & Over and 18 & Over Mixed are ESLs that start in the Fall as that is when courts are available since 40 & Over and 18 & Over will use all the available courts in the Winter and Spring.

The challenge is getting players rostered at the right level for these ESL teams, and what to do if a player is bumped up at year-end.

Say for example you are a very good 3.5, on your way to being bumped up to 4.0, and there is an ESL in your area for Adult 18 & Over that starts in September.
  • Should you be allowed to play as a 3.5 or should something be in place to let you know you are likely to be bumped up and must play as a 4.0 in the ESL?
  • And if you are allowed to play as a 3.5 and are bumped up at year-end, should you be allowed to keep playing as a 3.5?

A number of years ago, the USTA employed Early Start Ratings to get players rostered at the appropriate level.  The way these worked is that a snapshot was taken of your dynamic rating a month or two before the ESL was to start, and if you were above the bump threshold, you were given an "E" rating that bumped you up.  Now, these E ratings didn't show up on TennisLink, but were only on Early Start Rating lists that each section/district/area would publish.  But our hypothetical 3.5 above could very well be a 4.0E and have to play as a 4.0 in the ESL, thereby avoiding the situation where he is on a 3.5 team and now a 4.0 after year-end ratings come out.

Note that it was possible that our player would be a 3.5E and then be bumped up to 4.0 at year-end.  Because they were a 3.5E they would be allowed to continue playing on this team as a 3.5 despite the bump up, unless their rating reached the clearly above level mark.

Now, several years ago, the USTA decided to do away with Early Start Ratings as the feedback they received from players and sections was that they were confusing.  There is a point to this as they weren't published on TennisLink, every area has their own date to take the snapshot and publish ratings, there had to be a special process to deal with players appealing their E rating, etc.  It was also possible that a player would be an Early Start bump up, and then not be bumped up at year-end which confused some players.

In lieu of them though, players now simply play at their current year-end (or DQ'd/promoted) level in an ESL.  This is simpler on the surface, but introduces another type of confusion on the back-end, as the USTA also said that players from an ESL that are bumped up are not eligible to play at the lower level at Nationals.  But each section had the choice of how long a player could keep playing at the lower level prior to Nationals, and some sections promote them immediately on year-end ratings being published, while others let them play at the lower level through Sectionals.

So, depending on the section, our hypothetical 3.5 may be bumped up immediately and not be able to play on their 3.5 ESL team anymore, or would be able to keep playing on it, despite now being a 4.0, through 2019 Sectionals.

In those sections where players can't play through Sectionals or into 2019 at the lower level, year-end ratings bring a whole new significance.  A strong ESL team could be gutted by the bump ups and not be able to continue, or at least be significantly weaker than before.  This is arguably the right thing to do though as it is fair to all the other legitimate 3.5 teams that don't have now 4.0s playing for them.  Especially come Sectionals where there may be districts/states that didn't have an ESL and thus only have at-level players on their roster.  It really isn't fair to them to have to play a team from another district/state that has above level players now.

But even the teams that can stay intact through Sectionals may ultimately be impacted as their team could win Sectionals, and then lose players that are ineligible for Nationals, or even not be able to go to Nationals as they don't have the minimum roster to go.

I have no reason to believe we are going back to Early Start Ratings any time soon, but I wonder if having them was a better situation, and if the weaknesses of the system couldn't be fixed rather than throwing the whole system out.

Were Early Start Ratings confusing to you?  Or is the potential impact to teams losing players at some point more confusing and detrimental to a team?

Similarly, did Early Start Ratings create a more level and fair playing field for all?  Or because an E rating could still change at year-end it didn't really make a difference?

I personally liked having Early Start Ratings, but then again, I'm a bit of a propeller head that likes this stuff.  But I didn't think they were confusing and I thought they were more fair as players are more likely to be playing at the right level.  And it also avoided the situation we have today where districts/states without ESLs are at an inherent disadvantage to those with ESLs.

But what do you think? Vote in the poll below.


What do you think of Early Start Leagues / Ratings?

Early Start Ratings were confusing, glad they are gone
The new system is more confusing with different rules on eligibility and out of level players being rostered
The new system isn't fair to other teams in areas where players can keep playing after being bumped
Players should get to stay on the team through Nationals even if they are bumped up
Early Start Leagues should go away to solve the problem
Keep the new system, it is better than Early Start Ratings
Bring Early Start Ratings back
Check all that apply, but please don't choose answers that conflict with each other
Created with PollMaker

Sunday, November 18, 2018

So, USTA League Nationals are over, when will 2018 year-end NTRP ratings be published?

With the completion of Mixed Nationals this weekend, everyone's attention will now turn to year-end ratings so players can prepare for their 2018 leagues and form teams at the right level.  Naturally, to know the right level, players need to know their 2018 year-end rating.

First, players that played at least three matches in a league that counts towards ratings will receive a new year-end rating for 2018 (there are a few more qualifiers, but that is the basic criteria).  If you didn't play the requisite three matches, your prior rating will carry over, unless it is 3 years old in which case it will expire and you will have to self-rate again to join a team in 2019.

In some areas like mine, 2019 leagues start early in January and teams must form in December.  In fact, we have to have team forms with minimum rosters turned in the first week of December, so learning what rating one is at as early as possible is desirable.

The general rule of thumb is that ratings are published around December 1st, or another rule of thumb is the week after Thanksgiving.  These are usually close to the same, but this year Thanksgiving is very early, and December 1st is a Saturday.  Hopefully they don't want for the first Monday of December as that would be the 3rd and really put a squeeze on teams in my area having to turn forms in I believe on the 4th.

Here is a history of when they've been published the last few years, typically late in the day/evening PST which is sometimes early AM the next day EST:



We see a lot of Monday releases, but the last couple years they were mid-week.

Given that Thanksgiving is so early, my guess (and hope?) is that ratings are published mid/late the week after Thanksgiving, somewhere around the 29th.  Although, in 2012 when Thanksgiving was early, they were out early the week after.  I do not think they will wait for December though.  But who knows, my guess could be wrong.

Last year the USTA put a countdown timer up on TennisLink telling us when ratings would be published.  It will be interesting to see if they do that again.

We are less than two weeks from the ratings likely being published, so many of you will just wait to see where you end up, but if you are interested in a more detailed analysis showing how I estimate all your matches have rated and how they affected your dynamic rating, contact me for a report.

2018 40 & Over Mixed National Champions: Caribbean, Southern, and Missouri Valley times two

The 40 & Over USTA League Mixed Nationals were played in Orlando this weekend, and after what sounds like some good competition, everything finished up smoothly on Sunday and National Champions were determined.

At the 6.0 level, Caribbean took the title with a 3-0 win over Pacific Northwest.

The 7.0 level, where there had been a 7-way tie for three spots, found Southern playing Hawaii in the final and pulling out a 2-1 win.

At the 8.0 level, Missouri Valley took out Southern Cal 2-1.

And at the 9.0 level, a similarly sounding final happened with Missouri Valley beating Southern 2-1, one of the two wins being decided in a match tie-break.

Congratulations all!

And with all Nationals complete, our attention now turns to year-end ratings.  Stay tuned for more details in the next few days, or contact me if interested in a report estimating where your rating may be.

Saturday, November 17, 2018

Tie-breakers galore at 40 & Over Mixed USTA League Nationals - A 7-way tie for 7.0!

Back when the new format for USTA League Nationals was announced, I wrote about the possibilities it presented and what might happen, a key point being that there could be large ties between teams to determine who would advance to the semi-finals, and the tie-breakers used would be very important.

We've had several scenarios where teams have advanced to the semis based on the currently used tie-breakers that wouldn't have had the tie-breakers been, in my opinion, more equitable, and this weekend for the 40 & Over Mixed we've had some similarly interesting and/or very large ties and a couple cases things could have been different with different rules.

Below, I'll show the standings for the semi-finalists and ties showing each team's win/loss record, their individual courts won/lost, their sets won/lost, and games won/lost.  This is similar to what TennisLink shows and uses for breaking ties, but I've added the sets won and games won stats.  The official tie-breaker doesn't include these but I'm including for discussion purposes.

First, the 7.0 level had what may be the largest tie ever broken in USTA history with 7 teams tied at 3-1 for 2nd thru 8th.  Hawaii was 4-0, and then Southern, Texas, and Caribbean got the other spots having lost fewer sets than Mid-Atlantic.

TeamWLIWILSWSLGWGL
Hawaii4010221613180
Southern319318713082
Texas319319713099
Caribbean3193198136103
Mid-Atlantic319321913391
Midwest318417812196
Pacific NW3184181012893
Middle States3175171211495

But if we take a look at set differential and it were used, the order would change.  Texas and Mid-Atlantic were +12 and would have taken the #2 and #3 spots, and then Southern and Caribbean would have been +11, and with Southern having fewer sets lost, they would have taken the #4 spot.  But the USTA regulations do not look at sets won or sets differential so Mid-Atlantic is out.

It is also interesting to see that of the 7 teams tied, Caribbean lost the most games so had it gotten to that tie-breaker, they would not have advanced.

Next, the 6.0 level had three 4-0 teams, and then a 3-way tie at 3-1 for one spot.  PNW got the spot on fewer games lost compared to Midwest since they had the same number of sets lost.

TeamWLIWILSWSLGWGL
Southern4011123414071
Caribbean4011123414075
Northern409319713577
Pacific NW319318812382
Midwest3193198130113
Hawaii318417813180

We see though that while PNW and Midwest tied on sets lost, Midwest actually won more sets, and if sets differential was used first, they would have advanced.  But the USTA regulations ignore sets won or sets differential and just go straight to games lost after sets lost so PNW gets the spot.

The 9.0 level had one 4-0 team and a 4-way tie for three spots.  The last semi-finalist spot came down to Southern and Eastern tied all the way to sets lost, and Southern got the spot on fewer games lost.

TeamWLIWILSWSLGWGL
Missouri Valley4010220713395
Southern Cal3110221513272
Northern Cal3193191011297
Southern3184191013386
Eastern31841610121109

Had sets differential been used, Southern would have still gotten the spot with a +9 vs +6 for Eastern.

As far as head-to-head, Eastern gave SoCal their loss, Southern gave NorCal their loss, MoValley gave Southern their loss, and Texas (2-2) gave Eastern their loss.  Head-to-head is used after individual court wins/losses so didn't have a role here.

The 8.0 level was boring, two 4-0 teams in Missouri Valley and Southern Cal, and two 3-1 teams, although Intermountain did get the #3 spot over Southern having lost fewer games since they were tied on sets lost, and sets won for that matter.

So there was some very close and competitive tennis played so far, and a few situations that highlight the importance of fair and equitable tie-breakers.  But regardless of how the ties were broken, good luck to the semi-finalists tomorrow!

2018 USTA League 40 & Over Mixed Update

Part way into day 2, things are tight at all levels at the last 2018 Nationals being played in Orlando this weekend.

At the 6.0 level, as I write this a full 11 teams have one or fewer losses and are still in it.  And with just three undefeated teams, there is very likely to be a tie for at least one spot.  The updated simulation has Northern, Southern, and Caribbean advancing with PNW and Midwest in a tie for 4th.

The 7.0 level has 9 teams with one loss or fewer and just two undefeated.  The current simulation has a 6-way tie for 2nd behind an undefeated Hawaii, three teams will be left out if it plays out that way.

At 8.0, there are 9 teams with one or fewer losses and just three undefeated.  The simulation says just a 3-way tie at 3-1 behind SoCal and Midwest, but a large group at 2-2 right behind them.

And at 9.0, there are 7 teams with one or fewer losses and just one undefeated.  A big tie at 3-1 is possible but the simulation says just 4 tied for 2nd behind MoValley.

It will be fun to see how close these end up being.

Friday, November 16, 2018

When a DQ isn't a DQ?

Note: Apologies in advance to those that have regulations phobia.  I'm going to be doing some quoting and explaining of some USTA League regulations.

Note 2: If you read through it all, I do get around to asking for your opinion at the end, would love to get people's thoughts here on the blog or in Facebook comments.

Now, on to the topic at hand.

It is pretty common knowledge that a USTA League player that is 3-strike DQ'd and promoted to the next level is not allowed play at their lower/DQ'd level after the DQ occurs. Here is the rule from the USTA Regulations (emphasis mine).
2.04B(3) Following an NTRP Dynamic Disqualification, the player may not play at the disqualified NTRP level of play or any lower NTRP level of play in either singles or doubles for the remainder of the Championship Year and for the succeeding Championship Year.  In any Division using combined levels, the combined NTRP rating of the disqualified player and partner may not surpass the level of competition. A player who has been moved up as a result of an NTRP Dynamic Disqualification in the Adult Division must immediately adjust his/her NTRP level of play in the Mixed Division.

It is also pretty well known that a player cannot be DQ'd at Nationals.  Here is that rule:
2.04B(2) Championships NTRP Dynamic Disqualification. There will be no NTRP Dynamic Disqualifications at National Championships.

What is perhaps not well known and I've seen happen is that a player can be DQ'd after Nationals if they accumulate a third strike at Nationals.  Here is a relevant rule on how DQ's would affect play at subsequent championships/Nationals (emphasis mine).
2.04B(5) Individuals who receive their third strike while participating in an Early Start League (ESL) must immediately adjust their NTRP level of play. Such players may not participate at the disqualified NTRP level in a previous year’s championship for which they may have qualified.

Note that this seems go to against another rule saying strikes aren't even checked at Nationals (emphasis mine):
2.04B NTRP Dynamic Disqualification Procedures. Dynamic ratings will be calculated for all Adult Division players during local league competition and at every level of championship competition below National Championships to determine if any players have reached the NTRP Dynamic Disqualification criteria

But I have seen players DQ'd as a result of play at Nationals, so 2.04B only seems to be followed during the event but the strikes are calculated after the event.  In fact, it happened this year in an interesting case.

The player was self-rated this year and went to one of the Adult Nationals events in mid-October, and apparently got a third strike there (despite 2.04B) as they are shown at the next higher level with a 'D' now on TennisLink, and the date shown for the DQ is in October a few days after that Nationals finished with no other matches played.  Note that this player is rostered on a team in a 2019 ESL which is perhaps why their strikes were checked and the DQ/promotion recorded which is why I quoted 2.04B(5) above and believe it applies.

The player also went to 18 & Over Mixed Nationals three weeks after the Adult Nationals and per rule 2.04B(3) and 2.04B(5) above, would seem to have been required to play at their new/promoted level, but they played with a partner rated such that it is clear they were allowed to play at their lower/DQ'd level.

Further, the player also qualified for 40 & Over Mixed Nationals this weekend and the same thing happened again.

So why were they allowed to play?  The rules do seem to be at least a bit ambiguous, or at least somewhat in conflict with each other, so must be interpreted, and apparently the interpretation was to allow them to play.

One interpretation (A) of the rules is that DQ's don't happen at a given Nationals event, e.g. a player will not be DQ'd during the weekend of a specific event they are playing at, but can occur after the event like we see happened in this case.  Given the player shows up as a 'D' at the promoted level on TennisLink, this would seem to be what happened, the player was DQ'd.

Another interpretation (B) might be that once Nationals start, a player cannot be DQ'd from any Nationals event that year.  This is perhaps the interpretation used to allow them to play, even though the player does show up as DQ'd on TennisLink.  The rationale may be that the DQ should not have been calculated per 2.04B and so that supersedes it being published and they are acting as if it wasn't published.

I am not a lawyer, but my reading of the rules is that interpretation A is more correct, the player should not have been allowed to play at their lower/DQ'd level.  In my opinion, more of the rules support it, and more importantly the rules that say no DQ's occur or strikes aren't checked at Nationals don't seem to have been followed, the player does show up on TennisLink at DQ'd, so the other rules should then kick in.

Even if you consider the "at" in 2.04B(2) to be ambiguous to a specific event or all Nationals, 2.04B(5) applies in this case as the player is participating in an ESL, got their third strike, and so must immediately adjust their NTRP level, including not being able to participate at the DQ'd level in a 2018 championship.  The boxes for this rule seem to have all been checked and thus it should apply.  And 2.04B(5) would seem to have been written explicitly for this scenario.

If the intent of the "at" in 2.04B(2) was to mean cannot be DQ'd from any Nationals event once they have been started, and 2.04B(5) is there only to potentially DQ a player prior to any Nationals event starting, one can see that side of the argument.  But then why do they show up as DQ'd on TennisLink?

Playing devil's advocate for a minute, I can see an argument that a player may be playing at Nationals on consecutive weekends and have made travel plans for the second already when they are DQ'd after the first, and that it would not be fair to not allow them to play and be out the travel expenses.  I can poke two holes in this though.

First, this is the accepted danger/risk of having self-rated players on a Nationals team.  The USTA already gives a wide berth allowing self-rated players to improve and not get strikes.  By the time they've gotten 3 strikes they really have demonstrated themselves to be way above level and that their self-rating was not accurate.  And in this case, the player went 5-1 at Nationals at the new/promoted level!  Allowing them to play isn't fair to all their opponents and opposing teams, a significantly larger number than the player in question and their teammates.  Shouldn't the rule be biased towards being fair to the larger number of players, and fair to those that didn't violate a rule and get DQ'd?

Second, the events affected were Mixed where it being a "combo" league, it was possible for this player to still play at their promoted level on the team, just with a lower rated partner.  E.g. for 7.0 Mixed, a 3.5 that is DQ'd to 4.0 can still play on the team, they just must play with a 3.0 partner.  So this player was still eligible to play both Mixed events, just with a different partner, so their travel expenses would not be for nought.

Clearly there is some interpretation to be made with the rules and regardless of what the right interpretation was, one was made and the players at the Mixed Nationals this year will carry on.  What I think is more important now is that this is cleared up in the future, specifically that the language in the regulations is clarified or enhanced to be clear on how this should be handled, even if it is interpretation B, just so there isn't controversy about it.

But I've stated my case for not allowing them to play.  What do you think and why?  Am I being too strict it not allowing them to play?  Should there be some leniency during Nationals?  How should the regulations be changed or clarified?



Thursday, November 15, 2018

Weather Update: Getting _to_ USTA League Nationals!

Normally I give an update on the weather at the site for a Nationals weekend.  This weekend in Orlando the weather looks very good, 70s and sunny.  However, getting to Orlando may be the issue.  With a winter storm hitting the Northeast, I'm hearing of delayed and even cancelled flights.

I'm not sure what the contingency plan is if a team just can't make it.  They could just give all opponents default wins, or just not include the match in the standings, both have a different effect on how teams would be identified for the semis.

Ideally, if a team were to be missing, the schedule would get redone without the team so that all teams still play four matches.  With the schedule simulation work I've done, I know I could do that in minutes, not sure if the USTA is prepared to or not.

I do hope it doesn't come to that though.  Here's hoping everyone's flights go as scheduled or alternate plans can be found to get all the teams there on-time.

Safe travels everyone!

Wednesday, November 14, 2018

Simulating/Predicting the 2018 USTA League Nationals semi-finalists and champion - 40 & Over Mixed 8.0 (corrected)

I've written about my simulations of the new format for USTA League Nationals, and hinted at who some favorites are, but now it is time to actually use the simulations to predict who the semi-finalists will be and of those, who the favorite to win it all is.

This is the set of my simulations for Mixed, and because Mixed are combination leagues, it doesn't make sense to use a small number for the top average as that would unfairly give an advantage to teams with a lot of uneven pairs.  Also, some teams have players that play just Mixed and so don't have Adult ratings.  Thus, I'm using the top-12 average for each team using my estimated ratings, plus I'm using the Mixed ratings I can calculate rather than the normal Adult ratings, and then account for some variation in play by each team and run through a million iterations to see what the most common result might be.

Here is the 40 & Over Mixed 8.0 that is coming up this weekend in Orlando, FL.

There is the full 17 teams and with that and the schedule the USTA came up with, there is actually a chance of six(!) teams being undefeated.  That shouldn't be the case, it is entirely possible to have a schedule where no more than five teams can be undefeated.  Now, the chances are exceedingly low of six 4-0 teams occurring, 0.001%, but it is possible, and five teams undefeated has a 4% chance.

The teams with the best chance to finish in at least a tie for 4th are:
  • Southern Cal - 89%
  • Northern Cal - 81%
  • Midwest - 80%
  • Hawaii - 80%

Just the first above has a 4-0 record as most likely, and there is just four teams in a tie at 3-1 as the most likely record, Eastern being the other team.  Five teams are at 2-2 though and could get to 3-1.

If the teams predicted make the semis, Hawaii would be a narrow favorite over SoCal and NorCal over Midwest, with Hawaii the very narrow pick in the final.

If you are on one of the teams, or any team going to Nationals, it is not too late to get more details from my simulation or a flight report to scout opponents and plan your line-ups.  While the above is what the simulation says, the ratings aren't perfect between sections, players have good and bad days, not everyone will make the trip or players other than the top-12 will play, so some smart captaining can still go a long way and my reports can really help.  Contact me if interested in any reports.

Update: Made a correction due to marking a player ineligible which changed things a bit.

Simulating/Predicting the 2018 USTA League Nationals semi-finalists and champion - 40 & Over Mixed 7.0

I've written about my simulations of the new format for USTA League Nationals, and hinted at who some favorites are, but now it is time to actually use the simulations to predict who the semi-finalists will be and of those, who the favorite to win it all is.

This is the set of my simulations for Mixed, and because Mixed are combination leagues, it doesn't make sense to use a small number for the top average as that would unfairly give an advantage to teams with a lot of uneven pairs.  Also, some teams have players that play just Mixed and so don't have Adult ratings.  Thus, I'm using the top-12 average for each team using my estimated ratings, plus I'm using the Mixed ratings I can calculate rather than the normal Adult ratings, and then account for some variation in play by each team and run through a million iterations to see what the most common result might be.

Here is the 40 & Over Mixed 7.0 that is coming up this weekend in Orlando, FL.

There is the full 17 teams and with that and the schedule the USTA came up with, there is actually a chance of six(!) teams being undefeated.  That shouldn't be the case, it is entirely possible to have a schedule where no more than five teams can be undefeated.  Now, the chances are exceedingly low of six 4-0 teams occurring, 0.004%, but it is possible, and five teams undefeated has a 3% chance.

The teams with the best chance to finish in at least a tie for 4th are:
  • Northern Cal - 99%
  • Southern - 93%
  • Middle States - 91%
  • Hawaii - 70%

The first three above have a most likely record of 4-0, and Mid-Atlantic and New England join Hawaii with most likely records of 3-1.

If the teams predicted make the semis, NorCal would be favored over either Hawaii and Middle-States very narrowly over Southern, with NorCal the pick in the final

If you are on one of the teams, or any team going to Nationals, it is not too late to get more details from my simulation or a flight report to scout opponents and plan your line-ups.  While the above is what the simulation says, the ratings aren't perfect between sections, players have good and bad days, not everyone will make the trip or players other than the top-12 will play, so some smart captaining can still go a long way and my reports can really help.  Contact me if interested in any reports.

Simulating/Predicting the 2018 USTA League Nationals semi-finalists and champion - 40 & Over Mixed 6.0

I've written about my simulations of the new format for USTA League Nationals, and hinted at who some favorites are, but now it is time to actually use the simulations to predict who the semi-finalists will be and of those, who the favorite to win it all is.

This is the set of my simulations for Mixed, and because Mixed are combination leagues, it doesn't make sense to use a small number for the top average as that would unfairly give an advantage to teams with a lot of uneven pairs.  Also, some teams have players that play just Mixed and so don't have Adult ratings.  Thus, I'm using the top-12 average for each team using my estimated ratings, plus I'm using the Mixed ratings I can calculate rather than the normal Adult ratings, and then account for some variation in play by each team and run through a million iterations to see what the most common result might be.

Here is the 40 & Over Mixed 6.0 that is coming up this weekend in Orlando, FL.

There is the full 17 teams and with that, there is a quite high chance of five undefeated teams where one would miss out on the semis.  The simulation says a 7% of this and a whopping 37% chance of four undefeated.  This event appears to be pretty top heavy.

The teams with the best chance to finish in at least a tie for 4th are:
  • Mid-Atlantic - 99%
  • Caribbean - 91%
  • Southern - 86%
  • Hawaii - 81%

All the above four have a 4-0 record as most likely, but another team, NorCal, has a 40% chance of 4-0 too.  A few others with a decent chance to make the semis are Southwest and Northern.

If the teams predicted make the semis, Hawaii would be favored over either Mid-Atlantic and Caribbean over Southern, with Hawaii the pick in the final

If you are on one of the teams, or any team going to Nationals, it is not too late to get more details from my simulation or a flight report to scout opponents and plan your line-ups.  While the above is what the simulation says, the ratings aren't perfect between sections, players have good and bad days, not everyone will make the trip or players other than the top-12 will play, so some smart captaining can still go a long way and my reports can really help.  Contact me if interested in any reports.

Tuesday, November 13, 2018

Simulating/Predicting the 2018 USTA League Nationals semi-finalists and champion - 40 & Over Mixed 9.0

I've written about my simulations of the new format for USTA League Nationals, and hinted at who some favorites are, but now it is time to actually use the simulations to predict who the semi-finalists will be and of those, who the favorite to win it all is.

This is the set of my simulations for Mixed, and because Mixed are combination leagues, it doesn't make sense to use a small number for the top average as that would unfairly give an advantage to teams with a lot of uneven pairs.  Also, some teams have players that play just Mixed and so don't have Adult ratings.  Thus, I'm using the top-12 average for each team using my estimated ratings, plus I'm using the Mixed ratings I can calculate rather than the normal Adult ratings, and then account for some variation in play by each team and run through a million iterations to see what the most common result might be.

Here is the 40 & Over Mixed 9.0 that is coming up this weekend in Orlando, FL.

With just 14 teams, there is no chance of five teams finishing undefeated, and in fact there is just a 0.2% chance of four teams at 4-0.

The teams with the best chance to finish in at least a tie for 4th are:
  • Northern Cal - 98%
  • Southern Cal - 97%
  • Texas - 63%
  • Missouri Valley - 57%
  • Intermountain - 57%

The top-2 teams separate themselves and are the only teams with a most likely record of 4-0.  The three other teams plus Southern are expected to finish 3-1 but several 2-2 teams too.

If the teams predicted make the semis, NorCal would be favored over either  MoValley or Intermountain and SoCal over Texas, with SoCal the pick in the final

If you are on one of the teams, or any team going to Nationals, it is not too late to get more details from my simulation or a flight report to scout opponents and plan your line-ups.  While the above is what the simulation says, the ratings aren't perfect between sections, players have good and bad days, not everyone will make the trip or players other than the top-12 will play, so some smart captaining can still go a long way and my reports can really help.  Contact me if interested in any reports.

Monday, November 12, 2018

2018 18 & Over Mixed National Champions: Caribbean, SoCal, MoValley, Mid-Atlantic, Southern

The 18 & Over USTA League Mixed Nationals were played in Mobile this weekend, and after a little format change on Friday due to some rain, everything finished up smoothly on Saturday and Sunday and National Champions were determined.

At the 6.0 level Caribbean, won their final 2-1 over Florida to take the title.

The 7.0 level found Southern Cal taking the title 3-0 over Texas.

At 8.0, Missouri Valley beat Texas in their final 3-0.

The 9.0 level had Mid-Atlantic beat Southern 2-1, one court going to a match tie-break.

And at the 10.0 level, Southern beat Florida 2-1.

Five different levels, five different champs!  Congrats to all!

Saturday, November 10, 2018

USTA League 18 & Over Mixed Nationals Day 2 Update

Day 2 of 18 & Over Mixed Nationals in Mobile is complete and we know the semi-finalists.

It appears all matches were played using full scoring formats, thanks to getting most if not all the matches in yesterday despite some rain due to using modified Fast Four scoring to shorten the matches a bit.

At the 6.0 level, there were four 4-0 teams, Caribbean, Texas, Florida, and Southern, so no tie-breakers or controversy.  Although tie-breakers were used for seeding the teams and Florida got the #3 seed over Southern on fewer games lost despite Southern having a sets record of 21-8 vs 19-8 for Florida.  But those are the tie-breakers being used.

The 7.0 level had two 4-0 teams, then a 5-way tie for two spots at 3-1.  The teams advancing were SoCal, Middle States, Texas, and Southern.  Midwest and PNW lost out having won one fewer court so a pretty straight-forward tie-breaker.

At the 8.0 level, there were three 4-0 teams in Missouri Valley, Texas, and Florida, and New England was tied with PNW at 3-1 but won one more court to got the #4 spot.

The 9.0 level was a little more interesting.  There was one 4-0 team in Missouri Valley, but then five 3-1 teams for three spots.  Texas and Southern separated themselves with 10-2 records on courts so took seeds #2 and #3, but then one team needed to be picked from Mid-Atlantic, Northern, and SoCal.  The former two were 9-3 on courts and Mid-Atlantic lost two fewer courts, so they got the spot.  SoCal got left out at 8-4 on courts.

But a deeper look reveals that Texas, Mid-Atlantic, and Northern all played New England, who didn't win a court the entire weekend going 0-12.  So those teams benefited from an easy opponent that SoCal didn't.  A deeper look yet shows that each of those three teams also benefited from New England defaulting a court each match.  SoCal obviously didn't have the benefit of that advantage and had to play each of their courts.  I do not know why the New England was allowed to field a team with a roster of 5, but they very well may have lost the defaulted court had it been played anyway given they were a weaker team.  But this shows that strength of schedule very well could have played a role in who was able to advance.

The 10.0 level, four teams finished 3-1 and all advanced, Southern, Florida, Midwest, and NorCal.  We almost had a 2-2 team make the semis, but alas not.

So the finals are set, and the weather should be fine tomorrow.  Good luck to all!

Friday, November 9, 2018

18 & Over Mixed Nationals Day 1 Update

Day one of 18 & Over Mixed Nationals is complete, or nearly so, and the event is more or less still on track despite the rain that came late in the day.

The USTA made the right decision in my opinion to start the day using a modified Fast Four format in order to complete matches a little quicker, and it appears it worked.  As I write this, a few matches are still not recorded so may still being completed or just haven't been entered yet, but by and large, all the day one matches are complete meaning day two can proceed as planned.

While Fast Four is not ideal when everyone is planning on playing normal sets, I think it is better than the alternative of having a third of the matches still to play and the scheduling issues that would ensue.

And the competition is shaping up to be good.

The 6.0 level has a whopping seven teams still undefeated, and some of those play each other tomorrow so it can't stay that way, but it will be tight and could come down to tie-breakers to decide who advances.  There is just a 1.5% chance four finish undefeated, but a 6-way tie at 3-1 for 2nd thru 4th is likely.

The 7.0 level has five, perhaps six (one still to be recorded) undefeated so is similarly competitive.  But almost no chance of four undefeated, a 6-way tie at 3-1 for the 4 spots is likely.

The 8.0 level will have five undefeated and five at 1-1, so a lot of teams still in it.  But no chance of four finishing 4-0, a 4-way tie at 3-1 for three spots likely, but a 7-way tie at 2-2 behind means a bunch of teams could get to 3-1 and make it interesting.

The 9.0 level has six teams undefeated and just three at 1-1 and a 14% chance of finishing with four at 4-0.  There is a greater than 1% chance of five undefeated still.

The 10.0 level has just nine teams so just three undefeated meaning we are headed for some tie-breakers.  There is a very good chance a 2-2 team makes the semis.

Good luck to everyone tomorrow playing full scoring!


Simulating/Predicting the 2018 USTA League Nationals semi-finalists and champion - 18 & Over Mixed 6.0

I've written about my simulations of the new format for USTA League Nationals, and hinted at who some favorites are, but now it is time to actually use the simulations to predict who the semi-finalists will be and of those, who the favorite to win it all is.

This is continuing my simulations for Mixed, and because Mixed are combination leagues, it doesn't make sense to use a small number for the top average as that would unfairly give an advantage to teams with a lot of uneven pairs.  Also, some teams have players that play just Mixed and so don't have Adult ratings.  Thus, I'm using the top-16 average for each team using my estimated ratings, plus I'm using the Mixed ratings I can calculate rather than the normal Adult ratings, and then account for some variation in play by each team and run through a million iterations to see what the most common result might be.

Here is the 18 & Over Mixed 6.0 that is being played this weekend in Mobile, AL.  I am posting this late, but it is using information prior to competition starting.

Remarkably, there is a chance of six undefeated, although it only occurred once in the million simulations.  Even with that, there is just a 0.01% chance of five and 0.4% chance of four undefeated.  That means there is a good chance of a tie for fourth and it would likely have 4 of 5 teams in it.

The teams with the best chance to finish in at least a tie for 4th are:
  • Mid-Atlantic - 85%
  • Southern Cal - 67%
  • New England - 66%
  • Midwest - 56%

Those four teams separate themselves a bit, but just the top team has a most likely record of 4-0.  A whopping eight teams have most likely records of 2-2 led by NorCal and Southwest the most likely to make the semis.

If the four teams predicted make the semis, Mid-Atlantic would be favored over Midwest and SoCal over New England, with Mid-Atlantic the pick in the final.

If you are on one of the teams, or any team going to Nationals, it is not too late to get more details from my simulation or a flight report to scout opponents and plan your line-ups.  While the above is what the simulation says, the ratings aren't perfect between sections, players have good and bad days, not everyone will make the trip or players other than the top-10 will play, so some smart captaining can still go a long way and my reports can really help.  Contact me if interested in any reports.

Play is underway at 18 & Over Mixed Nationals

Play is underway this morning and several match results have already been entered for the 18 & Over Mixed Nationals being held in Mobile.  Yea!

Yesterday's forecast of 40-60% chance of rain most of the day has turned into 0% chance this morning, but still a chance of scattered showers at noon and then 4:00 pm, so continuous play is perhaps not going to happen, but some good long windows to get matches in.

With any luck, the USTA being proactive and going to a short format to start will allow most of today's scheduled matches to be played, and with a good forecast the next two days, the event should finish smoothly.

Thursday, November 8, 2018

Simulating/Predicting the 2018 USTA League Nationals semi-finalists and champion - 18 & Over Mixed 8.0

I've written about my simulations of the new format for USTA League Nationals, and hinted at who some favorites are, but now it is time to actually use the simulations to predict who the semi-finalists will be and of those, who the favorite to win it all is.

This is continuing my simulations for Mixed, and because Mixed are combination leagues, it doesn't make sense to use a small number for the top average as that would unfairly give an advantage to teams with a lot of uneven pairs.  Also, some teams have players that play just Mixed and so don't have Adult ratings.  Thus, I'm using the top-16 average for each team using my estimated ratings, plus I'm using the Mixed ratings I can calculate rather than the normal Adult ratings, and then account for some variation in play by each team and run through a million iterations to see what the most common result might be.

Here is the 18 & Over Mixed 8.0 that is coming up this weekend in Mobile, AL.

First, there is just a 0.02% chance of five teams finishing undefeated, so I don't think an undefeated team will be sent home.  There is a 1% chance of four 4-0 teams though.

This one should be more competitive than some of the other levels this weekend, there isn't two teams running away with it.  The teams with the best chance to finish in at least a tie for 4th are:
  • Mid-Atlantic - 80%
  • Missouri Valley - 71%
  • Middle States - 69%
  • Pacific Northwest - 54%
  • Northern Cal - 53%

These are all still over 50% so they are favorites, but the most likely record for each team is just 3-1, and there are 7 teams behind with a 2-2 record most likely that could certainly get to 3-1.  Midwest and Texas are most likely to 3-1, but Northern, New England, and Southern have decent chances too.

If the teams predicted make the semis, Mid-Atlantic would be favored over NorCal or PNW, and Middle States narrowly over MoValley, with Middle States the ever so narrow pick in the final.

If you are on one of the teams, or any team going to Nationals, it is not too late to get more details from my simulation or a flight report to scout opponents and plan your line-ups.  While the above is what the simulation says, the ratings aren't perfect between sections, players have good and bad days, not everyone will make the trip or players other than the top-10 will play, so some smart captaining can still go a long way and my reports can really help.  Contact me if interested in any reports.

Breaking News: 18 & Over Mixed Nationals going to short sets due to weather

I wrote a few days ago that the weather forecast for Mobile for this weekend's 18 & Over Mixed Nationals didn't look good for Friday,  an 80% chance of thunderstorms, and the current forecast still calls for scattered thunderstorms and 60% chance of precipitation.

Looking closer, the hourly forecast says a 40-60% chance of rain nearly all day which might make it difficult to get many matches in as any gaps in the rain may allow the courts to dry, right before it starts to rain again.

I'm guessing as a result of this forecast, I'm hearing that the USTA is preemptively going to a shorter scoring format from the start to get as many matches in as possible on Friday, and then be able to catch up on Saturday if needed.

I have not heard specifics, but I'm guessing it will be a format like what was done in Arlington using a modified Fast 4 with sets to 4 with no-ad, a set tie-break at 4-4, and either a set tie-break at 3-3 or just first to 4, and a match tie-break for the third set, perhaps just to 7 points instead of 10.

It is unfortunate it comes to this, I really was hoping the forecast would improve and the rain would clear out sooner, but this is perhaps the best that can be done given the site selection.

I will post more or update this as I learn more.

Update: Sets 1 and 2 will have a standard set tie-break at 4-4.  Here is the text sent to captains in an e-mail:

We will be playing best two out of three short sets with NO AD scoring and a 10 point coman tiebreak rotation if teams split sets. Teams will play to the first to 4, win by 2. If the set reaches 4-4 teams will play a 7 point coman tiebreak.