Monday, December 30, 2019

How many ways can you skin the "Points Per Position" cat?

Points per position (PPP) is suddenly becoming more popular as more sections adopt it in reaction to the 40 & Over division going to a 4-court format.

I've started taking a look at what points are most commonly used, and it appears there is no standard and there are a wide variety of allocations used.  It appears each section/district is given leeway to do what they see fit, and opinions vary widely.

Here is a list of all the options I've seen for 5-court formats:
  • 1,1,1,1,1 - Eastern, Midwest, New England
  • 4,3,2,5,3 - Middle States
  • 5,4,5,4,3 - Middle States
  • 5,4,6,4,3 - Eastern, Middle States
  • 5,6,4,3,2 - Eastern
  • 5,6,5,4,3 - Eastern, Middle States
  • 6,4,6,4,3 - Eastern
  • 6,5,6,5,3 - Eastern
  • 6,6,5,4,3 - Eastern

That is quite a collection.  But there are quite a few for 4-court formats too:
  • 1,1,1,1 - Eastern, Mid-Atlantic, New England
  • 1,2,1,1 - Northern California, Pacific Northwest
  • 4,3,5,4 - Middle States
  • 4,4,3,2 - Eastern
  • 4,5,4,3 - Middle States
  • 5,4,6,4 - Eastern
  • 5,5,4,3 - Middle States
  • 5,6,4,3 - Eastern
  • 6,6,4,3 - Eastern
  • 6,6,5,3 - Eastern
  • 10,6,4,2 - Texas

And for 3-court formats:
  • 1,1,1 - Eastern, Midwest, New England
  • 2,2,1 - Eastern
  • 3,2,3 - Middle States
  • 3,4,2 - Middle States
  • 4,3,2 - Eastern, Middle States
  • 5,4,3 - Middle States
  • 5,6,4 - Eastern
  • 6,4,3 - Eastern
  • 6,6,3 - Eastern
  • 7,4,3 - Middle States

Now, it is worth noting these allocations span different court formats, not all 5-court formats are 2S/3D, nor are all 4-court formats 1S/3D, etc.  But there is still a lot of variations, no one seems to be honing in on the best allocation to use.  And my image above is woefully under representing all the allocations out there!

What do you think?  What is the preferred allocations to use?

Thursday, December 26, 2019

Northern district of Eastern section to use Points-per-Position for 40 & Over for 2020 USTA League

News continues to trickle in on how the new 40 & Over 4-court format is being adopted across the USTA sections.  I've written about what NorCal and PNW have planned, and now I've heard what will be done in the Northern district of Eastern.

The news from there is that they will play 4-courts and be another adopter of points-per-position assigning 4 points for court 1 singles and doubles, 3 points for court 2 doubles, and 2 points for court 3 doubles.  Or 4432 in the shorthand I am using.

Since how this is being handled varies so much, I have started a page to capture what I learn so it can be a resource for all.  Please contact me if you have information to fill in any gaps.

Update on the tie-breakers for USTA League 4-court formats: Home team gets the win? 😕

Those that have been following my blog recently know that the USTA went to a 4-court format for 40 & Over for 2020.  The result of this is that 2-2 ties are possible and while there are some documented tie-breakers, they won't always break ties.  Yes, a match could end in a 2-2 tie and no winner be declared.

This has led some sections like Pacific Northwest and Northern California to adopt the points-per-position format for standings for the 40 & Over division so that teams just accumulate points and team wins aren't a factor in standings.  Other sections that used a 4-court format in the past elected to call the match a tie at this point and they had to handle standings manually as TennisLink can't show ties.

Now, above I said a match could end in 2-2 tie with no winner.  I've just learned that that is not entirely true!

Apparently, while the documented rules stop at Game Winning Percentage as the last tie-breaker (which is a meaningless tie-breaker), TennisLink has been programmed to decide the winner based on who is the home or visiting team!

I am told that originally the visiting team would be given the win, the idea being that they were able to tie the match despite having the disadvantage of playing at the other team's facility.  However, at some point recently, it changed to give the win to the home team!

Yes, you heard right, not only does the home team get the advantage of playing at home, but should the match end in a 2-2 tie, they get the benefit of the doubt and will get the win.   Further, consider a playoff match where who is home or visitor is arbitrary and that luck of the draw will determine the winner it seems.

I guess giving the win to the home team is deterministic, and avoid a coin flip, but it still seems wrong and if one was going to do this, giving it to the visitor makes more sense to me.  I also see more merit in treating it as a tie.

Of course points-per-position avoids this, well, until two teams are tied on points and breaking that tie goes to the head-to-head tie-breaker ...

Anyway, beware of the ties and know how your section handles it as all don't do the same thing, and the rules used may not be documented.

What do you think?

Update: I have looked deeper for examples of ties and what happened, and I've seen situations where there home team is a 2-2 winner and the visiting team is a 2-2 winner.  So it may be that there are local rules that determine how this is handled.  The lesson learned is that if you are in a 4-court league that isn't using points-per-position, ask your Local League Coordinator before the season starts what the tie-breaker is so rules aren't being made up as the season goes along.

Tuesday, December 24, 2019

Merry Christmas from the USTA? Denied appeals now granted

The Great Southern Bump of 2019 resulted in a lot of men from the Southern section being bumped up when year-end ratings came out.  Some of these bump ups were frankly very questionable, it seems the USTA did a blanket adjustment to several levels of players, and naturally quite a few players felt the bump up wasn't appropriate and tried to appeal back down.

Most of what I heard is that there were very few appeals down granted, but yesterday I heard from several folks that e-mails from the USTA had been sent to some players with the following message:
We recently found a problem with our auto-appeal system. In our investigation of the issue, we found that you may have attempted an appeal and may have been impacted. The problem has been resolved, and if desired, your automatic appeal can be attempted again.
At least some of these players appealed again and this time it was granted!

I have no idea what the problem was or what was done to fix it, but it appears just in time for Christmas, the USTA is granting appeals that were not originally granted.

So, if you had your appeal denied, perhaps you can try again, and if it is now granted, please let me know with a comment here or on Facebook or by dropping me a note at ratings@teravation.net.  The more reports I get, the more I can do to try to figure out a pattern or what happened.

Merry Christmas!


Sunday, December 22, 2019

NorCal to use Points-per-Position for 40 & Over for 2020 USTA League

After National introduced a 4-court format for 2020 40 & Over, there were bound to be changes and a month ago, Pacific Northwest announced they would be using Points-per-Position scoring for regular season standings, and now I've learned Northern California will be using PPP as well.

It was almost certain to happen, a 4-court format with the possibility for 2-2 ties just begs for a change to the format, and I hypothesized National was secretly pushing for PPP to be used more, and whether that is the case or not, it is happening.

Like PNW, NorCal is assigning one point to all courts other than court 1 doubles which will be assigned two points.  There will thus be five points up for grabs in each match, and the team match "win" has no meaning as standings are simply determined by accumulated points.

What do you think?  What other sections are introducing PPP for 40 & Over in 2020?

Monday, December 16, 2019

New local playoff format debuts in PNW / Seattle - Is it more equitable than before?

The Seattle area of the Pacific Northwest section just held their local playoffs for 18 & Over Mixed this past weekend.  With it, a new format for local playoffs was used that is interesting and worth discussing.

As background, in the Pacific Northwest section, our sectionals are generally eight teams from four to six districts.  Alaska is a district and often doesn't send a team, and sometimes Southern Oregon doesn't send a team for a given division/level/gender.

To fill out the eight teams then, two to four wildcards are handed out to the 2nd place teams from the other districts.  These are generally done giving wildcards first to the districts with the most teams at a level.  The result is that Seattle and Portland always get a wildcard, and Eastern Washington and Southwest Washington often do as well.

This leads to an interesting situation at local playoffs in each district where with a normal single elimination draw, the finals are between two teams that already know they are both going to Sectionals and the final carries far less weight than it might otherwise.

To give the final more meaning and significance, a new format was used this year, and from what I can tell may be for other division's local playoffs for the 2020 league year.

In the past, the local playoff format has taken the top two teams from each sub-flight and created a draw where 1st place teams play 2nd place teams, or if there aren't a power of two teams, the top teams get a 1st round bye.

This year, for districts where there is a wildcard slot for Sectionals, the format has changed to have all the sub-flight 1st place teams in one draw, and the 2nd place teams in another draw.  The 1st place teams play their draw to conclusion and the winner is the local playoff champion.  But the loser of the final does not automatically get the wildcard berth.

Instead, the 2nd place team draw is also played to completion and the winner of that draw then faces the 1st place team draw final loser for the wildcard spot.

It is a little confusing, but you can sort of think of it like a "winner's" bracket for the 1st place teams where the winner advances and the final loser still gets an extra chance to advance, and a "loser's" bracket for the 2nd place teams where not only do you have to win your draw, but play an extra match after to get the wildcard.

Here is an example of what the draw looks like.


This is an interesting approach and I think has some pros and cons.

On the surface, there is a clear advantage to being a 1st place team in your sub-flight as you get to go to the "winner's" bracket and can lose in the final and still be able to get the wildcard if you win the subsequent wildcard match.  Correspondingly, there is a disadvantage to being a 2nd place team as there is no safety net should you use in your draw's final, and you have to play and win an extra match to get the wildcard.

But is it really an advantage to be in the "winner's" bracket?

Presumably the 1st place teams are the stronger teams, so it is arguably a tougher road to win the bracket, and the safety net of being able to lose is there only if you make the final.  Additionally, in a three or five sub-flight scenario, a 1st place team may still have to play an extra match if they are the 3rd or 5th seeded team.

Consider a scenario with five sub-flights as the draw above is for, where one of them had several strong teams.  The sub-flight winner, team A, may finish with one loss and be seeded 5th.  This means they have to play an extra match against the 4th seed in the "winner's" draw.  They will have to win two matches against other 1st place teams to make the final, and then win a third against another 1st place team, or should they lose, turn around and play a fourth match and win it.

Now, the 2nd place team, team B, in the same sub-flight may have also had just one loss and be seeded 1st or 2nd in the "loser's" bracket and not have to play the extra match as the 4th and 5th seeds in this draw are required to.  Their road to the wildcard match requires winning just two matches against 2nd place teams, where they may get to play a team just coming off losing in the "winner's" bracket final having to play a fourth match.

So was it really an advantage to be the 1st place team?  Both have to win three matches to advance to Sectionals, and two matches in a row to get to a match where it is win and advance, but team B gets to do it against 2nd place teams while team A has to do it against 1st place teams.  Is the failsafe team A has of losing at that point and having a 2nd chance to advance in a fourth match worth it?

Obviously, if you are one of the top-3 seeds in the "winner's" bracket, you don't have to play the extra match and just having to win one match to make the final probably is worth having to do so against a first place team.  But even in this scenario one could perhaps argue that being a top-3 seed in the "loser's" bracket has some advantages.

Now, clearly this does make the "winner's" bracket final meaningful where it wasn't before, and it does perhaps give the 2nd place teams a more realistic chance of advancing, but it does sort of stink for the 1st place team that has to play the extra match or face a top-2 seed early while the 2nd place teams get less formidable opponents.

It certainly introduces a new dynamic though and perhaps makes the entire event more exciting.

What do you think?  Do you like this new format and how it makes the final matches more meaningful?  Or does it unfairly put the lower seeded 1st place teams at a disadvantage?

Friday, December 6, 2019

2020 USTA League Nationals Dates and Locations - Similar to 2019

Taking a break from all the year-end stats and bump analysis, I came across a document that has the dates and locations for 2020 Nationals.  Yes, 2019 Nationals finished up less than 4 weeks ago, but it is time to think about 2020.

Important: While this document is on a USTA site, I have not seen it linked to on usta.com or TennisLink, so it is perhaps not final and subject to change.

Nevertheless, in an effort to get the latest news out to my readers, here is a summary.

First, the locations are the same as 2019, Surprise, Vegas, Oklahoma City, and Orlando.  It appears things went well enough at the new OKC site that they are going back.

Second, the schedule is generally the same, with 18 & Over happening first with 40 & Over overlapping but starting a week later and going a week longer, and 55 & Over near the end and Mixed wrapping the sequence up.  This has been pretty standard for a number of years.

On to the specifics.

Note: As of February 24th, the USTA Nationals page has a slightly different schedule than originally posted and it is updated below.

October 2-4
  • OKC - 18 & Over 3.0
  • Surprise - 18 & Over 3.5
  • Vegas - 18 & Over 5.0

October 9-11
  • Vegas - 18 & Over 2.5 women
  • Surprise - 18 & Over 4.5
  • OKC - 40 & Over 3.5

October 16-18
  • Vegas - 18 & Over 4.0 40 & Over 4.5+
  • OKC - 40 & Over 3.0
  • Orlando - 40 & Over 4.5+ 18 & Over 4.0
  • Surprise - 55 & Over 8.0

October 23-25
  • Orlando - 40 & Over 4.0
  • Surprise - 55 & Over 6.0 / 9.0

October 30-November 1
  • Surprise - 55 & Over 7.0
  • Orlando - 18 & Over Mixed 6.0 / 8.0 / 10.0 7.0 / 9.0

November 6-8
  • Orlando - 18 & Over Mixed 7.0 / 9.0 6.0 / 8.0 / 10.0

November 13-15
  • Surprise - 40 & Over Mixed 7.0 / 8.0
  • Orlando - 40 & Over Mixed all levels 6.0 / 9.0

Note the document says the last event is the 13th and 14th, I have to think that is a typo and it is the 13th-15th like I listed it.

The themes are pretty much the same, Orlando hosts Mixed, Surprise hosts 55 & Over, and the other events get spread around a bit.  In fact, while some dates changed, the locations for each event is nearly identical to 2019.  It appears just the 40 & Over 3.0 has moved from Orlando to OKC and 40 & Over 4.0 moved from Surprise to Orlando.  Every other event is at the same location as last year.

I would have thought they'd change things up a bit so anyone going back to the same event would get a new location.  But perhaps they figure with all the bumps up, few players will be back?

In any case, for those optimistic about their chances of qualifying, you now know the (perhaps tentative) dates!

Digging in to the Great Southern Bump of 2019 - Level distribution by state

I wrote yesterday about how men in the Southern section were bumped up at significantly higher rates than the national averages.  I also took a look at the distribution of players by level in Southern compared to national averages.

The logical next step is to drill into the data and see if whatever was done was done consistently throughout the section or if some states were more/less impacted.

As a reminder and baseline, here is the distribution of Southern men by level.


See the earlier posting for all the details, but key thing is that there were and are still more 3.5s than 4.0s, and there were nearly twice as many 3.0s as 4.5s but those counts are now basically equal.  Nationally there are about the same 3.5s and 4.0s and roughly the same 3.0s and 4.5s.

Alphabetically, we will start with Alabama.


Alabama definitely had 2.5s and 3.0s bumped up, and given the influx of them there were 3.5s bumped up too, although just a modest drop in count.  There were a lot of 4.0s bumped up as the count went down despite the influx and the 4.5s were the beneficiaries, that count went way up as did the 5.0s.  There are now significantly more 4.5s than 3.0s, but still more 3.5s than 4.0s.

Next, Arkansas.


Arkansas did have players bumped up from 2.5 through 3.5, those counts did go down from 2018 to 2019, but 4.0s and above had their counts go up, supportive of the hypothesis that there was a general redistribution towards the higher levels.  Note that Arkansas still has fewer 4.0s than 3.5s but a lot closer than before, and it is nearly equal between the 3.0s and 4.5s.

Next, Georgia.


Georgia was only modestly out of whack with the national averages, but has almost no 2.5s now and 3.0s were bumped up too.  More 3.5s were bumped up than came in and while there were a lot of bumps up to 4.5, the 4.0s grew.  Georgia is now very close to the national averages with about the same number of 3.0s as 4.5s and 3.5s as 4.0s.

Next, Kentucky.


Kentucky had bump ups at 2.5 and 3.0, but a lot at 3.5 given their count went down even with the bump ups coming in 4.0 was the beneficiary with a big increase.  4.5 and 5.0 also went up significantly.  There are now significantly more 4.0s than 3.5s and 4.5s than 3.0s.  Kentucky was perhaps a state that did not need an adjustment to align with national averages, but still got it.

Next, Louisiana.


Louisiana also had bump ups, there are hardly any 2.5s anymore, and we see an increase in the number of 4.0s, 4.5s, and 5.0s.  They are not yet to the national averages with way more 3.0s than 4.5s and noticeably more 3.5s than 4.0s.

Next, Mississippi.


Mississippi had a lot of bump ups at 3.0s and the number of 3.5s went up as a result.  There were bump ups to 4.0 but more went from 4.0 to 4.5 and 4.5 to 5.0.  Mississippi does have nearly as many 4.5s as 3.0s, but there are still a lot more 3.5s than 4.0s.

Next, North Carolina.

North Carolina had a fair number of bump ups from the lower levels, the 4.5s and 5.0s having the largest increase.  They have more 4.5s than 3.0s now, and just a few fewer 4.0s than 3.5s.  They have more 4.5s than 3.0s, but still a few more 3.5s than 4.0s.

Next, South Carolina.


Just a small percent of 2.5s bumped up, but quite a few 3.0s went to 4.0, with a lot of 3.5s to 4.0, and 4.5 and 5.0 both picked up some players.  They are still biased towards the lower levels with a lot more 3.0s than 4.5s and a lot more 3.5s than 4.0s.

Last, Tennessee.


There are some remarkable changes here.  2.5s are non-existent, and nearly all 3.0s were bumped up to 3.5 where the count of 3.5s nearly tripled!  There had to be some bump downs from 4.0 for that to happen and the 4.0s stayed about the same while the 4.5s and 5.0s here took precipitous drops.  Tennessee is clearly an exception to the general bump up across the levels trend.  This probably deserves more research on its own.

What do you think?  Does this dovetail with what you've seen with your friends and teammates?

Thursday, December 5, 2019

The Great Southern Bump of 2019 - USTA Southern men get historic bump up rates

I've been going through my post-NTRP ratings published analysis and what was a little smoke when looking at bump up/down rates at a high level has turned into a raging fire as I've drilled in.

Here is a summary of what I've posted so far regarding bump up rates of players that had a C rating at the end of 2018 and 2019.
  • 2.5 men - Southern had 75% bumped up, nationally 52%
  • 3.0 men - Southern had 33.1% bumped up, nationally 21.8%
  • 3.5 men - Southern had 18.2% bumped up, nationally 11.2%
  • 4.0 men - Southern had 17.8% bumped up, nationally 10.9%
  • 4.5 men - Southern had 19.8% bumped up, nationally 6.7%
  • 5.0 men - Southern had 6.3% bumped up, nationally 1.5%

The Southern men had the highest bump up percentage among all the sections at every level!  And it was significantly above the national average, usually 60+% greater, but about 3 times greater for 4.5s and 4 times for 5.0s.  It is pretty much unheard of for there to be nearly 20% of 4.5s bumped up and anywhere close to 6% of 5.0s.

That begs the question, what has this done to the distribution of players across the levels?  Let's look again at the distribution nationally for the men.


In this chart, it shows stats for both 2018 and 2019 to easily see the change year to year, and the length of the bar is the number of players at that level, while the label at the end of the bar is the percentage of players at that level that year.

We see a "normal" distribution, the center being just about right between 3.5s and 4.0s.  And compared to 2018, not a dramatic change, the percent being 3.5 dropped slightly, 4.0s down half a percentage point, and the 4.5s up just over 2 points, with 5.0s up half a point.

Now, the distribution in Southern.


This is somewhat different in that the center of the "normal" curve is closer to 3.5 than 4.0.  We see that it has changed quite a bit though and moves closer to the mid-point between them with the 2.5 percent dropping 1 point, 3.0 percent dropping over 4 points, 3.5 dropping 0.4, 4.0 going up nearly a point, 4.5 going up a whopping 3.2 points, and 5.0 going up nearly 2 points.

Or put another way, it appears 2.5s moved to 3.0, more 3.0s moved to 3.5, but even with the influx of 3.5s the number went down slightly as just as many 3.5s moved to 4.0.  The number of 4.0s did grow, but there were still a lot that moved up to 4.5, and quite a few 4.5s moved to 5.0.

So there was a definite shift upwards which dovetails with the bump up percentages reported above.

One might ask the question, why was and is Southern biased more towards lower levels, and was the large number of bump ups a reaction to that to try and make the distribution closer to what it is nationally?  It could be that the USTA League playing population is older in Southern and happens to have more lower rated players, but clearly it is different than the national distribution, even after the noticeable bump ups.

However, an adjacent section, Florida, perhaps also with an older population of USTA League players, is actually biased slightly towards the higher rated players already.  Here is Florida's distribution chart.


Florida had some movement from 3.5 to 4.0 and 4.0 to 4.5, but already had more 4.0s than 3.5s, and more 4.5s than 3.0s unlike Southern.

So it would seem someone decided that Southern players were generally underrated, or at least out of whack with the national distribution, with a distribution biased too far towards 3.0 and 3.5, and steps needed to be taken to change that distribution and get more 4.0s and 4.5s.

What do you think?  Is the distribution of Southern's men being towards the lower levels an issue and was the redistribution needed and appropriate?

Note: "The Great Southern Bump of 2019" was a phrase I thought I coined, but an acquaintance had done so as well so credit goes to "OTL".

Wednesday, December 4, 2019

Analyzing 2019 USTA NTRP year-end ratings - Distribution of players by level

Next in the sequence of analyzing the 2019 year-end ratings, we'll take a break from bump up/down percentages and look at how many players there are at each level.

Should anyone want to see how this looked last year, take a look.

Here are the number of players, men or women, with C ratings at each NTRP level.  I've included 2018 year-end counts as well for comparison.  The percentages on each bar are the percent of the total at that level.


For those that have followed in past years, we see the "normal" distribution we expect.  Comparing to 2018, we see numeric increases at 2.5 thru 3.5 as well as 4.5 and 5.0.  Just 4.0 has a decrease in total number.  Note that while the 3.5 total count went up slightly, the percentage of the total dropped 0.1%.

Here is the distribution for the women.


This chart has a similar shape, but is biased more towards 3.0 than 4.0 and smaller numbers at 4.5 and above too.  Modest increases at 2.5 and 3.0, the rest pretty close to last year.

And for the men.


The men's shape is more biased towards the higher levels with nearly as many 4.0s as 3.5s.  Interestingly, there is a modest increase of 3.5s, 4.5s and 5.0s, but a decrease of 4.0s.  The increase of 4.5s is quite noticeable, and for those that have been following along, is likely due to the "Great Southern Bump of 2019".

Now of course, these are numbers across all sections.  Stay tuned for some drill downs into some of the sections, particularly Southern.

Analyzing 2019 USTA NTRP year-end ratings - Section bump rates by level and gender - 5.0 men and women

Next in my series of posts analyzing the 2019 year-end ratings I continue breaking down the bump rates by level and gender.  I already looked at bump rates for each section by gender, bump rates by level nationally, and started the by level analysis with the 3.5 men and women4.0 men and women3.0 men and women, and 4.5 men and women.

Next up the 5.0 levels.  First the women.


As you might expect, very few bumps up, only 7 sections really had any and quite a few had over 20% bumped down.

Next the men.


Here we see a lot of bump downs and very few bump ups, but yep, Southern is the exception with the fewest bump downs, less than 4% where most others were over 10% and many over 20%, and Southern had more 5.0s bumped up, over 6%, than down.

Stay tuned, more to come.

Analyzing 2019 USTA NTRP year-end ratings - Section bump rates by level and gender - 4.5 men and women

Next in my series of posts analyzing the 2019 year-end ratings I continue breaking down the bump rates by level and gender.  I already looked at bump rates for each section by gender, bump rates by level nationally, and started the by level analysis with the 3.5 men and women4.0 men and women, and 3.0 men and women.

Next up the 3.0 levels.  First the women.


As you might expect after seeing the 3.0 thru 4.0 analysis and how the bump ups go down as the levels go up, there are a lot more bump downs at 4.5.  What is interesting is that Caribbean is the section with the most bump ups at over 8%.  Only Middle States is close at just under 6%, everyone else is at 4% or lower.

Next the much awaited 4.5 men.


We see a similar trend, more bumps down at this level than the lower levels, except for Southern that had nearly 20% of 4.5s bumped up!  That is a higher percentage than the 4.0s bumped up to 4.5 and the 3.5s bumped to 4.0.  No other section was even close with only Texas at about 7.5% having noticeably more bump ups than down.  Mid-Atlantic who won 4.5 Nationals was about an even split.

I have no explanation for this.  IMHO the bump up rates for Southern at every level, but especially 4.5, can't be simply from normal year-end calculations and adjustments.  There seems to have been an agenda to adjust and/or redistribute things in Southern for some purpose for 2020.

Stay tuned, more to come.

Tuesday, December 3, 2019

Analyzing 2019 USTA NTRP year-end ratings - Section bump rates by level and gender - 3.0 men and women

Next in my series of posts analyzing the 2019 year-end ratings I continue breaking down the bump rates by level and gender.  I already looked at bump rates for each section by gender, bump rates by level nationally, and started the by level analysis with the 3.5 men and women and 4.0 men and women.

Next up the 3.0 levels.  First the women.


We see a lot more bumps up, every section has more up than down, but Caribbean leads the way at nearly 19% and Middle States, Northern, SoCal, and NoCal are all at 15% or higher.  Just Intermountain had 5% bumped down, the rest less.

This makes sense, 3.0s generally have more upside, so you would expect more bumps up.

Next the 3.0 men.


There is an even larger disparity between the bumps up and down, hardly any bumps down in any section, and Southern had nearly 33% of 3.0 men bumped up and Texas nearly 28%.  This is clearly a big reason why these two sections led the way across all levels.

But four other sections were over 20% in Caribbean (actually a fraction more than Texas), Northern, New England, and Southwest.

Stay tuned, more to come.

Analyzing 2019 USTA NTRP year-end ratings - Section bump rates by level and gender - 4.0 men and women

Next in my series of posts analyzing the 2019 year-end ratings I continue breaking down the bump rates by level and gender.  I already looked at bump rates for each section by gender, bump rates by level nationally, and started the by level analysis with the 3.5 men and women.

Next up the 4.0 levels.  First the women.


What jumps out is there are a surprising number of sections that have more players bumped down than up.  Caribbean had more the twice as many bumped down than up, but Midwest had the highest percentage at nearly 8%.  Only 5 sections had more bumps up than down, Pacific Northwest the most at over 8%, and then Southwest, Northern, and Eastern.

Next the 4.0 men.


We see the general trend of Southern and Texas having a large number of bump ups continue, Southern nearly 18%, Texas nearly 15% bumped up.  Next closest is Florida and Intermountain.  Only Caribbean and Hawaii had significantly more bumps down than up.  PNW also did narrowly, and NorCal barely had more up than down but both very lost percentages.  It appears the USTA thought NorCal and PNW had fairly rated 4.0 men with such little movement.

Stay tuned, more to come.

Analyzing 2019 USTA NTRP year-end ratings - Section bump rates by level and gender - 3.5 men and women

Next in my series of posts analyzing the 2019 year-end ratings I begin breaking down the bump rates by level and gender.  I already looked at bump rates for each section by gender, as well as bump rates by level nationally.

To start, I'll look at the 3.5 level starting with the women.


Caribbean and Southern Cal had the most bump ups, Mid-Atlantic not far behind, but these all under 10% and all of the rest under 7.5%.  No section had more bump downs than up, but Intermountain was close.

And next the men.


Here we see the trend general trend of Southern and Texas having the most bump ups applying to the 3.5 level too, Southern at over 18% and Texas nearly 15%.  Southern Cal, Northern, and Missouri Valley were each over 10% as well.  Only Southwest had more than 5% bumped down.

Monday, December 2, 2019

Analyzing 2019 USTA NTRP year-end ratings - Section bump rates by gender

2019 year-end ratings have been published, and my analysis has begun.

I already took a look at bump rates by level comparing to past years, now on to some bump rates by section and gender.

Here are the overall bump rates by section.


We see some typical bump rates of 8-10% up and 4-5% down, but there is some variation by section.  Southern and Texas lead the way with bump ups with 13% and 11% respectively, Caribbean and Middle States at 10.2% and Mid-Atlantic at 10%.  Hawaii had the fewest bump ups at just over 7%, very close to their bump downs.  Caribbean joined Hawaii with the most bump downs but still under 7%.

But it is more interesting to look at things by gender too.  Here are the women.


There are no big outliers, but Caribbean, Mid-Atlantic, Middle States, and Northern are all just over 10%.  Hawaii and Missouri Valley had the fewest bump ups but still over 7%.

The men had more variation, here they are.


We see that Southern and Texas are the big outliers, Southern having over 21% of men bumped up!  Texas at nearly 14% would be high in most years as well.

With a section as large as Southern, a lot of bump ups numbers wise is expect, that this is an awfully high percentage and would seem to indicate some kind of adjustment.  This has happened in past years with other sections, Caribbean was last year with an astounding 45% bumped up, and Caribbean and PNW were over 20% in 2014, so Southern and to a lesser degree Texas have been singled out this year.

This mirrors the anecdotal evidence I'm hearing, but it will be interesting to split it out by level too.  Stay tuned for that.

2019 USTA NTRP bump rates by level - Interesting Tennis League Stats

The USTA has published their 2019 year-end ratings and since I had just posted general bump up/down rates by level, I thought I'd update the table with the latest bump data.

What this table is showing is the players at a given level that were bumped up.  Specifically for example, 7.6% of 3.5s that were 2017 year-end 3.5s were bumped up to 4.0, and 4.1% were bumped down.  So the new data added is now showing that 8.1% of 2018 year-end 3.5s were bumped up to 4.0 and 4.0% were bumped down.

Here is the full table.

2019201820172016
Level   Up   Down   Up   Down   Up   Down
2.532%0%35%0%32%0%29%0%
3.014%3.2%14%3.1%14%2.9%14%3.1%
3.58.1%4.0%7.6%4.1%7.2%4.0%7.8%3.7%
4.07.2%5.6%4.7%5.9%4.3%5.7%4.8%5.2%
4.55.0%7.7%4.4%8.2%3.9%8.3%3.6%7.8%
5.01.4%20%1.9%17%1.6%17%1.7%17%
5.50%37%0.7%39%0%36%0%45%

The most noticeable change is that the bump up rate for 4.0s is significantly higher than past years, 4.x for three years but 7.2% for 2019.  That is a huge jump nationally and across genders and I'm sure there will be pockets where that is much higher (stay tuned for that).  The bump down rate for 4.0s was not significantly lower though.

4.5s also had a higher bump up rate but not nearly as much as the 4.0s.

Based on this, it would seem the USTA may have thought there was a glut of players at the top of the 4.0 range and did a little adjusting up, and perhaps also a bit with 4.5s to get a few more 5.0s.

I will work on all my normal analysis in the coming days, but wanted to get this out quickly.  Also, I'll look at slicing this data by gender and section/state too.

Update: Bump rates by section and gender here.

Sunday, December 1, 2019

2019 USTA League 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 12/31/2019 date so at least some are now available.  I think things showed up right at midnight ET, so technically on 12/2 as expected.  I just expected it later in the day on 12/2.

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 12/31/2019 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.

What USTA League players will be bumped up/down Monday when ratings are published?

The USTA League season wrapped up back in November and now that we've had Thanksgiving, it is time for year-end NTRP ratings to be published!

Word is that ratings should be published on Monday December 2nd, my guess is towards the end of the day, and a lot of people will be interested to see where they end up.  Some players had good years and are anticipating (or afraid of?) being bumped up, others had middle of the road seasons and expect to stay the same, while others perhaps had losing seasons and are afraid (or hoping?) for a bump down.

My ratings, while not perfect, are a very good indicator of where a player's rating will end up, and I can do still do reports for player's wanting to see where they will end up, or looking for an explanation of why they ended up where they did.

But, we can also look historically at some statistics of how past years have ended up to get an idea of what will happen this year.

What follows is a table showing bump down/up percentages by year and level for the past few years.

201820172016
Level   Up   Down   Up   Down   Up   Down
2.535%0%32%0%29%0%
3.014%3.1%14%2.9%14%3.1%
3.57.6%4.1%7.2%4.0%7.8%3.7%
4.04.7%5.9%4.3%5.7%4.8%5.2%
4.54.4%8.2%3.9%8.3%3.6%7.8%
5.01.9%17%1.6%17%1.7%17%
5.50.7%39%0%36%0%45%

We see the percentages are quite consistent for most levels, but in general there have been a slight increase in bump ups the past two years, and the bump downs have also had a small increase at several levels.

But given how consistent the bump percentages have been, it is probably pretty safe to assume 2019 will show very similar bump percentages as shown above.

Stay tuned, I'll be posting more stats and observations once year-end ratings are published.

Update: Updated the above table in this post adding in 2019 for comparison.