The USTA's flawed tie-breakers rear their ugly head again

[schmke]

The tie-breakers the USTA uses for standings in TennisLink, and thus used at Nationals, involve breaking ties between teams ultimately on sets lost and games lost. It has been this way for years.

This is fundamentally broken as it ignores sets won and games won, and when at Nationals teams may not even play each other and have the ability to hang lost sets/games on the other team, it is especially egregious.

Unfortunately these flawed tie-breakers have kicked in numerous times before (https://tt.tennis-warehouse.com/ind...e-breakers-at-nationals.628794/#post-12786020) and they happened again this weekend.

The 55+ 8.0 women had two 4-0 teams and then four 3-1 teams as follows:

MS - 3-1 / 9-3 / 19-7
NE - 3-1 / 8-4 / 17-9
SO - 3-1 / 8-4 / 19-9
MA - 3-1 / 7-5 / 15-11

I've listed the teams in the order shown on TennisLink but why is New England ahead of Southern? Middle States has a better court record than the others so they get the 3rd spot, and Mid-Atlantic has the worst court record so they get 6th, but 4th and 5th are confusing.

The issue is that the USTA's tie-breakers only look at sets lost, not sets won. So NE and SO are tied and it goes to games lost where NE had lost fewer. But shouldn't sets won have been considered where SO won two more than NE?

The USTA says no, which is effectively telling SO, who was able to extend courts they lost to 3rd set TBs and thus won additional sets, that doing so did nothing for them.

The reason this is important at Nationals is that teams very well may not play each other, so you need the criteria used to break ties to be as equitable as possible. If one team happens to play a tougher schedule and lose a team match because of that, they shouldn't also be penalized by a flawed tie-breaker like this when being compared with a team with a weaker schedule. The fact that the team with the stronger schedule was able to extend matches and win extra sets should be considered.

In this case, Southern's loss was to Florida, who was 4-0 / 10-2 and the top seed. Southern gave them one of their court losses and so losing to Florida was certainly a "quality loss". Who did New England lose to? 1-3 Hawaii. That is not a quality loss.

Also, Southern beat Hawaii 2-1 who New England lost to, and Southern beat MOValley 3-0 while New England only beat them 2-1. So by any common opponent measure, Southern would be ahead too.

Southern's opponents had a record of 6-10 while New England's were 4-12. Neither exactly tough schedules, but Southern's noticeably tougher.

So, IMHO, the USTA got it wrong and it is so easy to fix. Come on USTA, get it right!

 

[travlerajm]

Not defending the current system. But one could argue that putting more value on games lost would help deter strategic tanking.

 

[schmke]

Pretty sure that isn't the reason. If this is how the USTA combats tanking, they aren't even using a pea shooter.

And using games lost is broken even worse than sets lost.

Consider a match where the sets are close but a team loses 7-5,7-5. They lost 14 games. If they had been completely outmatched and lost 6-0,6-0, that would have been better as it is only 12 games lost. What? How does that make sense?

 

[travlerajm]

Pretty sure that isn't the reason. If this is how the USTA combats tanking, they aren't even using a pea shooter.

And using games lost is broken even worse than sets lost.

Consider a match where the sets are close but a team loses 7-5,7-5. They lost 14 games. If they had been completely outmatched and lost 6-0,6-0, that would have been better as it is only 12 games lost. What? How does that make sense?

That’s a good point. Set win % and game win % would make a lot more sense.

 

[schmke]

That’s a good point. Set win % and game win % would make a lot more sense.

Which is what I proposed with a formal regulation change request year ago and they said no.

 

[Cashman]

I expect the reason for deliberately omitting games/sets won is to prevent teams who play a lot of timed matches being unduly disadvantaged. When you start taking won sets/games into account, there is suddenly a big difference between winning a match 6-2, 6-3, or winning it 6-2, 4-3 because time ran out.

Now of course, these things cut both ways and there is no guarantee that if a timed match played out that you would be a net beneficiary. There are a number of games/sets left on the table - some of them you will win, and some of them you will lose. And it is also true that under the current tiebreak rules, teams who play a lot of timed matches do receive a bit of a benefit - from the fact that they have undoubtedly avoided conceding at least some games/sets that would hurt them in a countback.

But - we have to consider that tiebreak rules are of greatest importance to the better teams in a league (because those are the teams that are in the running for playoffs, podiums etc.). And for a top-table team that wins most of their matches, you could reasonably expect them to win a lot more of the unplayed games/unfinished sets than they would lose. My hypothesis (perhaps an analysis opportunity @schmke) is that if you looked at a cohort of teams with a similar match win % across the country, there would be a definite correlation between set/game win % and how many timed matches they played. This has the potential to create some quite inequitable outcomes if you start leaning on metrics incorporating won games/sets as a tiebreaker.

So in the context of some top teams playing a lot of timed matches and other top teams playing a lot of untimed matches, I do think that focusing on what you lost rather than what you won makes a certain amount of sense.

Now I assume Nationals matches are all played out - which means the above is all a bit moot for that tournament (at least in isolation). But it is also true that tiebreak rules are invariably poorly understood and there is a good argument for keeping them as consistent as possible. I imagine you might get a bunch of very confused and angry players if they unexpectedly got dudded by a Nationals tiebreak rule that was different to the one in play for the rest of the year.

 

[am1899]

Which is what I proposed with a formal regulation change request year ago and they said no.

I’d tell you how shocked I am. But that would a lie. It’s astonishing how many good ideas I’ve heard about that USTA turned up their nose at. They’re either incompetent or they’re so sure they know everything that they’re blind. (Maybe some of both). God help us.

 

[schmke]

I expect the reason for deliberately omitting games/sets won is to prevent teams who play a lot of timed matches being unduly disadvantaged. When you start taking won sets/games into account, there is suddenly a big difference between winning a match 6-2, 6-3, or winning it 6-2, 4-3 because time ran out.

Now of course, these things cut both ways and there is no guarantee that if a timed match played out that you would be a net beneficiary. There are a number of games/sets left on the table - some of them you will win, and some of them you will lose. And it is also true that under the current tiebreak rules, teams who play a lot of timed matches do receive a bit of a benefit - from the fact that they have undoubtedly avoided conceding at least some games/sets that would hurt them in a countback.

But - we have to consider that tiebreak rules are of greatest importance to the better teams in a league (because those are the teams that are in the running for playoffs, podiums etc.). And for a top-table team that wins most of their matches, you could reasonably expect them to win a lot more of the unplayed games/unfinished sets than they would lose. My hypothesis (perhaps an analysis opportunity @schmke) is that if you looked at a cohort of teams with a similar match win % across the country, there would be a definite correlation between set/game win % and how many timed matches they played. This has the potential to create some quite inequitable outcomes if you start leaning on metrics incorporating won games/sets as a tiebreaker.

So in the context of some top teams playing a lot of timed matches and other top teams playing a lot of untimed matches, I do think that focusing on what you lost rather than what you won makes a certain amount of sense.

Now I assume Nationals matches are all played out - which means the above is all a bit moot for that tournament (at least in isolation). But it is also true that tiebreak rules are invariably poorly understood and there is a good argument for keeping them as consistent as possible. I imagine you might get a bunch of very confused and angry players if they unexpectedly got dudded by a Nationals tiebreak rule that was different to the one in play for the rest of the year.

Good point and something interesting to perhaps research.

But you are right, Nationals plays everything out, and unless there are weather delays, that includes playing out deuces and not even playing no-ad.

Now, I say that, but sometimes a team will retire a court when they know they've won the team match already, especially in the final round when they know they won't need the retired court. But that is different than a timed match and the team is hurting themselves if they retire so that is on them and they wouldn't (you'd think) do that if it was going to be meaningful.

 

[Chalkdust]

That’s a good point. Set win % and game win % would make a lot more sense.

Nah, that doesn't work, I mean it's not like it's the accepted tiebreaker procedure used by round robin tournaments like the Atp Tour Finals or anything...

 

[socallefty]

I heard SoCal won 55+ 8.0 Men this weekend. Are they mostly computer ranked players or do they have a lot of self-rated players?

 

[schmke]

I heard SoCal won 55+ 8.0 Men this weekend. Are they mostly computer ranked players or do they have a lot of self-rated players?

They had two 3.5S guys who both should be bumped up to 4.0 at year-end. One went 33-0 including 6-0 in 18+ and 5-0 in 40+ and was arguably about to be (or perhaps will be now this week) DQ'd. So perhaps there is a self-rate problem even for 55+ players!

 

[schmke]

Starting a new thread to discuss the SoCal 55+ 7.0/8.0 teams: https://tt.tennis-warehouse.com/index.php?threads/it-happens-in-55-too.738696/

 

[JLyon]

well it is the USTA, they could not hit water if they fell off a boat.
Orlando swamp needs to be drained for new leadership

 

[Pass750]

Such great expertise being shared on these threads, thank you. I don’t play USTA, but after reading all these threads I don’t think I ever would! USTA sounds like a terribly run organization. I am happy playing local leagues where the people are playing at the right levels.

These threads do provide great entertainment value though.

 

[sweetrugger]

Such great expertise being shared on these threads, thank you. I don’t play USTA, but after reading all these threads I don’t think I ever would! USTA sounds like a terribly run organization. I am happy playing local leagues where the people are playing at the right levels.

These threads do provide great entertainment value though.

I think it depends on what you are trying to get out of it. I live in California, where due to population density, there are a number of districts relatively close together. Some people really love playing tennis and competing, so they play on multiple teams each season. Some folks want a drinking team that plays tennis on occasion...that's possible too! Each season there are personality clashes, but a vast majority of matches are positive, fun, etc.

 

[cks]

I don’t play USTA, but after reading all these threads I don’t think I ever would! USTA sounds like a terribly run organization. I am happy playing local leagues where the people are playing at the right levels.

Like you, I enjoy reading the USTA league stories on this site. And we recently had quite a few threads on the teams that managed to make it through Sectionals and on to Nationals.

Before you disregard USTA league, I would recommend you try it out first. I just started playing USTA league this year (January 2022), and so far it has been a positive experience for me. I've met new folks who share my passion for this game and had a chance to play at different tennis clubs located around my city. Even though there may be some out-of-level players or captains who may recruit out-of-level players that make it to Nationals, I'm still glad I joined and played USTA league this year. I think these stories/complaints/sandbagging make up a small portion of the total amount of tennis matches played under USTA league. Just my 2 cents.

Best of luck if you decide to try USTA league.

 

[J_R_B]

Which is what I proposed with a formal regulation change request year ago and they said no.

Between this and the 4 court format in 40+, perhaps the USTA is intentionally trying to be as stupid as possible in the decisions they make? To what end, I don't know, but there certainly is empirical evidence for it.