Florida comes up with potentially the worst use of unflighted round-robin yet

schmke

schmke

Legend
I wrote yesterday and today about a scenario that could happen at the Florida Sectionals for 18+ 9.0 Mixed this weekend, where it is possible that using the unflighted round-robin format, five teams could finish undefeated, ... but only two advance to the final.

Go find my blog (sorry, I'm not allowed to post links here) to read the details, but this diagram illustrates the issue. It shows the teams and who plays who.

2022-FL-18-90X.png


You can see they took 10 teams, split them into two pods, and then had each team play four teams in the other pods. Nice and easy to do, but, ...

What if all the teams in one pod win? You have five 5-0 teams! Uh-oh.

Normally you'd think that the unflighted round-robin would work great with just 10 teams and each team playing four matches, after all Nationals has 17 teams playing four matches each and it works (but it has its pitfalls too), but what Florida is doing here is the absolute worst way to do unflighted round-robin.

One wonders why they didn't just do two flights of five with the flight winners playing the final ...
 
travlerajm said:
Maybe they did it because it looks cool on a diagram.
Click to expand...
ROTFL, had they been savvy enough to create the diagram, the problem should have jumped out and grabbed them!

But you are right, someone probably did this diagram to do the schedule and it simply never occurred to them there was a problem.
 
I am pulling for Sarasota (SRQ) and Santa Barbara (SBA) if anyone from SRQ or SBA advances to any tennis events. :). Why? two of my favorite overnights.

www.visitsarasota.com

Five Stops for Tasty Tacos in Sarasota County

www.visitsarasota.com www.visitsarasota.com

santabarbaraca.com

Santa Barbara Taco Trail | Visit Santa Barbara

Stumbling upon restaurants and taquerias serving up mouth-watering tacos and authentic Mexican fare is dangerously easy to do here in The American Riviera®.
santabarbaraca.com santabarbaraca.com

East Beach Tacos for the win! ;)
 
schmke said:
One wonders why they didn't just do two flights of five with the flight winners playing the final ...
Click to expand...
What had happened was...
They were intending to do exactly that - two flight of five, each team plays the other four teams in the flight.
So flights "A" and "B:, with A1 vs A2, A1 vs A3, etc, etc.
But then as someone was typing it up, they got confused and wrote it up as A1 vs B2, A1 vs B3, etc.
And then by the time the mistake was noticed, things had already been set in motion...
So then a quick about face to say it was done this way intentionally, because reasons.
Stranger things have happened...
 
I can think of one thing to say in favor of this format, which is that if there are more than two undefeated teams, they will all be from the same pod, which means they will have played mostly the same opponents. So at least the tie-breaker is comparing matches won/lost against common opponents. There are other formats where an undefeated team getting sent home played an entirely non-overlapping set of opponents than an advancing team, which is arguably worse?

I'm not saying that this format is good, just food for thought.
 
TennisOTM said:
I can think of one thing to say in favor of this format, which is that if there are more than two undefeated teams, they will all be from the same pod, which means they will have played mostly the same opponents. So at least the tie-breaker is comparing matches won/lost against common opponents. There are other formats where an undefeated team getting sent home played an entirely non-overlapping set of opponents than an advancing team, which is arguably worse?

I'm not saying that this format is good, just food for thought.
Click to expand...
You are right, it does make the tie-breakers used more equitable than they would be if each team played completely different opponents, but each one will still have played one different opponent. Good point/observation though.
 
I was looking at the formats the USTA supports for tournaments a while back and couldn't really fathom why you would have "round robin" formats like this.

Round Robin(where each team/player actually plays all others), group play, and swiss are all legitimate formats. This version of round robin is worse version of swiss. It doesn't work if you don't pair based on standings at some point.
 
ponzicrash said:
I was looking at the formats the USTA supports for tournaments a while back and couldn't really fathom why you would have "round robin" formats like this.

Round Robin(where each team/player actually plays all others), group play, and swiss are all legitimate formats. This version of round robin is worse version of swiss. It doesn't work if you don't pair based on standings at some point.
Click to expand...
This format has advantages in certain situations.

If you have "odd" (not just numerically odd) numbers of teams it can work better than alternatives. One of, if not the, first use was for Nationals where you have 17 (nominally) teams. You can't have flights of the same size with 17 teams, and Nationals for years had three flights of four teams and one flight of five. The five team flight got an extra match, but also had to play an extra match making the winner more tired for the semi potentially. The unflighted round-robin format allows getting all teams four matches which is a good thing. If you have numbers like 13, 11, 9, 7 it can also work well.

It also is a format where a team with a bad draw can have a shot at still making the semis. Consider the old way, only one team from each flight advances, even if the 2nd place team in a flight is better than all the other flight winners. With unflighted round-robin, they can lose to a really good team but still win their other matches and make it to the semis.

Now, there are challenges with the format if not done right. You have to have each team play enough matches to reasonably assure yourself that you won't be sending undefeated teams home before the semis or final, and ideally you take steps to help ensure the teams have reasonably similar schedules, otherwise when you apply the tie-breakers it isn't really fair to the team who played the tougher schedule. You also need to not use broken tie-breakers like the USTA still does.

Now, doing what Florida did with 10 teams total is just silly because having two flights of five would have gotten all the teams four matches just like they are playing as constructed. But even with the format, they could have avoided the potential issue I highlighted if they hadn't done the schedule the way they did.
 
schmke said:
If you have "odd" (not just numerically odd) numbers of teams it can work better than alternatives. One of, if not the, first use was for Nationals where you have 17 (nominally) teams. You can't have flights of the same size with 17 teams, and Nationals for years had three flights of four teams and one flight of five. The five team flight got an extra match, but also had to play an extra match making the winner more tired for the semi potentially. The unflighted round-robin format allows getting all teams four matches which is a good thing. If you have numbers like 13, 11, 9, 7 it can also work well.
Click to expand...

If the "odd" number of teams is the only thing keeping them from doing a sensible flight or group format, then why not consider inviting more teams to Nationals? They could invite the 2nd place team from some Sections, perhaps from the biggest sections (to award participation) or to the Sections that produced the top teams from the prior year. Or maybe just offer a series of "play-in" tournaments around the country, where any team that advanced to Sectionals (but didn't win) could earn one of the wild-card spots for Nationals.
 
TennisOTM said:
If the "odd" number of teams is the only thing keeping them from doing a sensible flight or group format, then why not consider inviting more teams to Nationals? They could invite the 2nd place team from some Sections, perhaps from the biggest sections (to award participation) or to the Sections that produced the top teams from the prior year. Or maybe just offer a series of "play-in" tournaments around the country, where any team that advanced to Sectionals (but didn't win) could earn one of the wild-card spots for Nationals.
Click to expand...
Interesting idea, and not that it couldn't be done, but there are probably lots of obstacles including politics between sections, having enough time for any "play-in" tournament (some sections are just finishing their Sectionals now with Nationals starting in two weeks), the logistics of play-in tournaments between sections and the travel and costs involved, rules for who is eligible, but also adding teams meaning an expansion of the Nationals event to more teams/matches and complications from the logistics of that and requiring facilities/sites to have sufficient courts, or having to change Nationals from a 3 day to 4 day event, etc.
 
One "good pod" team lost today, but the other four are all 2-0. My simulation says a 55% chance we end up with three or four teams all at 4-0.
 
schmke said:
Interesting idea, and not that it couldn't be done, but there are probably lots of obstacles including politics between sections, having enough time for any "play-in" tournament (some sections are just finishing their Sectionals now with Nationals starting in two weeks), the logistics of play-in tournaments between sections and the travel and costs involved, rules for who is eligible, but also adding teams meaning an expansion of the Nationals event to more teams/matches and complications from the logistics of that and requiring facilities/sites to have sufficient courts, or having to change Nationals from a 3 day to 4 day event, etc.
Click to expand...

Here's the simplest, fairest way I can think of off-hand: take the 17 Sectionals winners as usual, then add 3 wild card teams, one each from 3 different Sections chosen randomly well before the league year starts. Rotate through all the Sections in subsequent years so every Section gets a wild card spot every 5-6 years. Sections can decide for themselves how to choose two teams during their designated wild card years.

Then with your 17+3=20 teams, make 4 groups of 5 teams. Play a full round robin in each group, so four guaranteed matches for every team, just like it is now. Then you have 4 group winners for the semi-finals and finals.

This adds 6 matches per tournament compared to now, which I suppose adds up to a lot across all the levels male and female. Too much? Any other drawbacks to this idea?
 
TennisOTM said:
Here's the simplest, fairest way I can think of off-hand: take the 17 Sectionals winners as usual, then add 3 wild card teams, one each from 3 different Sections chosen randomly well before the league year starts. Rotate through all the Sections in subsequent years so every Section gets a wild card spot every 5-6 years. Sections can decide for themselves how to choose two teams during their designated wild card years.

Then with your 17+3=20 teams, make 4 groups of 5 teams. Play a full round robin in each group, so four guaranteed matches for every team, just like it is now. Then you have 4 group winners for the semi-finals and finals.

This adds 6 matches per tournament compared to now, which I suppose adds up to a lot across all the levels male and female. Too much? Any other drawbacks to this idea?
Click to expand...
I like the idea.

The additional 6 matches isn't insignificant, especially when it is really 12 since you have both genders. For 18+ with 5 courts, that is an additional 60 courts that will be required.

An advantage of the unflighted round-robin that I think I mentioned earlier is that in your format, a bad draw getting in a tough flight with one or two other top teams dooms all but one from advancing, while the unflighted round-robin still gives a team that loses a shot. You can of course argue that the team that can't win in their flight shouldn't have an argument to advance because they'd have to beat the strong team eventually anyway, but unflighted round-robin may get the "best" 4 teams in the semis while your format just gets the best team from each flight, not necessarily the same thing.

Last, at some levels every section doesn't send a team and you end up with 16, 15, or even fewer teams. The USTA can't know this in advance so they can't know how many wildcards there will be well before the league year starts. They can of course simply have a wildcard order and if they need more than three, go down the list to get them, but that would have to be done at some levels.
 
Well, it happened. Four teams went 4-0 and two will not advance unless they ad lib and do semis (top-2 to the final was the format).

Miami-Dade was 11-1 on courts and gets first.

Sarasota and Hills were both 10-2 on courts but Sarasota gets second due to losing one fewer set. That means Hill goes home despite going 4-0 / 10-2

Broward was 4-0 / 9-3 and in fourth also won't advance.
 
Defaults are normal for a few reasons, some that you describe.

Having multiple levels at the same event where a player can be in both is certainly problematic as you described.

But even when it is just a single level being played there are times a default will occur on the ongoing courts once a team has the team win in order to have those players save energy for the next round. And when this is unflighted round-robin, that default may benefit the opponent should they be in the running to advance should it come down to tie-breakers. And of course even if there aren't defaults, the teams that are out of it will often roll out weaker line-ups to get everyone a match, or just go through the motions which can also affect standings and tie-breakers.

But it is no different in pro sports when teams that are out of it may be directly or indirectly tanking to get better draft position, or playing backups to get them experience. Or even teams that have their playoff spot wrapped up sitting starters to protect/rest them.
 
schmke said:
An advantage of the unflighted round-robin that I think I mentioned earlier is that in your format, a bad draw getting in a tough flight with one or two other top teams dooms all but one from advancing, while the unflighted round-robin still gives a team that loses a shot. You can of course argue that the team that can't win in their flight shouldn't have an argument to advance because they'd have to beat the strong team eventually anyway, but unflighted round-robin may get the "best" 4 teams in the semis while your format just gets the best team from each flight, not necessarily the same thing.
Click to expand...

It's not clear to me that the unflighted round-robin would have the advantage here, because it suffers from the same potential problem of a top team not advancing to the semis because of a bad draw. In an extreme case, you could even have the BEST team get sent home b/c of a bad draw: if the best team gets four tough matchups and beats all of them 3-2 or 4-1, while there are four other undefeated teams that benefit from 5-0 wins against much weaker teams. This worst-case scenario would never happen in the group format. That particular example might be unlikely, but it seems feasible that the 3rd or 4th "best" teams regularly fail to advance in the current format.

Of course the group format would be improved by some kind of seeding ala FIFA World Cup to avoid extreme groups of death, but I'm not sure USTA would want to attempt something like that.
 
I wonder if it would be better in general to reduce the number of matches in a round-robin phase in favor of adding another round to the elimination tournament phase.

For example for Nationals, have a 3-match round robin followed by an elimination tournament of the top 8 teams. Essentially you are replacing the 4th round-robin match with the quarterfinals of the tournament for those teams that make it. The teams that don't make the top 8 could get a fourth match against another non-advancing team. Those would obviously be somewhat meaningless, but the same thing happens in the current format with 0-3 or 1-2 teams playing their last match. With an odd total number of teams, one team (the last-place 0-3 team?) would get left out of having a fourth match, but maybe that team wouldn't really mind checking out early after getting blown out 3 times...
 
TennisOTM said:
I wonder if it would be better in general to reduce the number of matches in a round-robin phase in favor of adding another round to the elimination tournament phase.

For example for Nationals, have a 3-match round robin followed by an elimination tournament of the top 8 teams. Essentially you are replacing the 4th round-robin match with the quarterfinals of the tournament for those teams that make it. The teams that don't make the top 8 could get a fourth match against another non-advancing team. Those would obviously be somewhat meaningless, but the same thing happens in the current format with 0-3 or 1-2 teams playing their last match. With an odd total number of teams, one team (the last-place 0-3 team?) would get left out of having a fourth match, but maybe that team wouldn't really mind checking out early after getting blown out 3 times...
Click to expand...
I like the idea, but two gotchas.

One is that odd team out. A perk of the unflighted round-robin was getting every team four matches, and your suggestion works for that, except for the odd team out when there are the full complement of 17 teams (or 15 or any other odd number).

But the bigger problem is that when you have 17 (or any other odd number) you can't have everyone play exactly three matches. With 17 teams and 4 matches, that is a nice 34 team matches that have to be scheduled and played. The even number of matches allows you to deal with the odd number of teams. When you play just three matches though, the math doesn't work. You have one team that only plays two matches, or one team that plays four.

That said, with unflighted round-robin and the challenges of differing schedules and the flawed standings tie-breakers, I do like the idea of going deeper and taking 8 teams to the quarters. That decreases the risk that a deserving team gets left out.
 
schmke said:
I wrote yesterday and today about a scenario that could happen at the Florida Sectionals for 18+ 9.0 Mixed this weekend, where it is possible that using the unflighted round-robin format, five teams could finish undefeated, ... but only two advance to the final.

Go find my blog (sorry, I'm not allowed to post links here) to read the details, but this diagram illustrates the issue. It shows the teams and who plays who.

2022-FL-18-90X.png


You can see they took 10 teams, split them into two pods, and then had each team play four teams in the other pods. Nice and easy to do, but, ...

What if all the teams in one pod win? You have five 5-0 teams! Uh-oh.

Normally you'd think that the unflighted round-robin would work great with just 10 teams and each team playing four matches, after all Nationals has 17 teams playing four matches each and it works (but it has its pitfalls too), but what Florida is doing here is the absolute worst way to do unflighted round-robin.

One wonders why they didn't just do two flights of five with the flight winners playing the final ...
Click to expand...
Wait, so the USTA came up with the absolute dumbest way to handle a critical playoff?!? LMAO, color me shocked.
 
schmke said:
I like the idea, but two gotchas.

One is that odd team out. A perk of the unflighted round-robin was getting every team four matches, and your suggestion works for that, except for the odd team out when there are the full complement of 17 teams (or 15 or any other odd number).

But the bigger problem is that when you have 17 (or any other odd number) you can't have everyone play exactly three matches. With 17 teams and 4 matches, that is a nice 34 team matches that have to be scheduled and played. The even number of matches allows you to deal with the odd number of teams. When you play just three matches though, the math doesn't work. You have one team that only plays two matches, or one team that plays four.

That said, with unflighted round-robin and the challenges of differing schedules and the flawed standings tie-breakers, I do like the idea of going deeper and taking 8 teams to the quarters. That decreases the risk that a deserving team gets left out.
Click to expand...
If you expand it to 8 advancing and a QF round, now you're almost always going to have a massive 2-2 tie for those last spots where some teams will be in and others out, although at 2-2, you really don't have an argument that you were screwed. Using the unflighted round robin as is, the only adjustment I would MANDATE is some kind of play-in to the semis when there are >4 undefeated teams. The level of injustice of sending an undefeated team home because of a ridiculous, meaningless tiebreaker is too much, but if you lost somewhere, even just once to a superteam, well, you should have won if you wanted to advance. In the case of this farce of playoff in FL, that would mean expanding to a 4 team semifinal instead of just a 2 team final.

I prefer the flighted approach. The idea is to crown the best champion, not to get the best 4 semifinalists. If you want to be in the semis, win your flight. If you're the second best team in the tournament and you're randomly in a flight with the best team, that's just bad luck, but you weren't going to win anyway. To me, excluding an undefeated team that may have had a tougher schedule but didn't lose runs a MUCH greater risk of missing the true champion than excluding a team that lost once that might be the second best team in the tournament.
 
schmke said:
But the bigger problem is that when you have 17 (or any other odd number) you can't have everyone play exactly three matches. With 17 teams and 4 matches, that is a nice 34 team matches that have to be scheduled and played. The even number of matches allows you to deal with the odd number of teams. When you play just three matches though, the math doesn't work. You have one team that only plays two matches, or one team that plays four.
Click to expand...

Whoa, had not thought of that problem! How about this idea for the tricky 17-team scenario:

Make 3 groups of size 5, 6, and 6. The size-5 group plays a full a round robin, and the size-6 groups play a near-full round robin (4 matches for each team). Then take the 3 group winners and the best group runner-up to the semis. To break ties within a group, use head-to-head results if applicable, followed by match record against common opponents only.

A size-6 group might have 2 undefeated teams at 4-0 who didn't play each other, but they will likely both advance. In the rarer case where both size-6 groups have 2 undefeated teams, well, that's a flaw, but the same thing could happen in the current format when there are five 4-0 teams, right?

I think I like this better than the current format because most of the tiebreakers (the ones that occur within a group) are fairer, because you can use head-to-head and or common opponent results. However you will likely have to compare 3-1 teams across different groups (who have zero common opponents) to get the 4th semifinalist. But selecting the 4th team can be iffy in the current format as well, whenever there are less than 4 undefeated teams.
 
Does USTA Florida not have a regional play off before the sectionals like USTA Eastern does? Having a regional/district playoff would halve the number of teams that go to sectionals, thereby allowing every team to play every other team in the sectionals.
 
You must log in or register to reply here.

Similar threads

schmke
SoCal trying to one-up NorCal with inappropriate use of unflighted round-robin
Replies
5
Views
1K
esgee48
esgee48
B
Round round and "a half"??
Replies
14
Views
898
TennisOTM
T
T
Delayed DQ?
Replies
8
Views
903
schmke
schmke
ncgator
USTA State Championship Rainout
Replies
38
Views
3K
struggle
S
schmke
New playoff format in situations where top-2 teams advance
Replies
9
Views
2K
schmke
schmke
Back
Top