Upsets the Rule in this Tournament

[Fairfield Fan]

I count 8 games in the first round in which a lower seed beat a higher seed. In the second round, I count 5 games with upsets, out of a total of sixteen games.

If you take out the top seeds, that means that in the second round nearly half (5 out of 12 games) resulted in upsets. In the first round, it meant that, if you take out the first three seeds, since they were clearly paired with inferior opponents, 8 of 20 games resulted in upsets. For both rounds, the seedings meant that the results could have been virtually replicated by throwing darts a a board with the teams' names on it.

So it seems that the NCAA seeding committee has hardly done a job that suggests competence. Would it be better to seed just the first four against the last four? And then just draw names out of a hat for all other games? I just can't believe that so many "upsets" are occurring. Seems more likely that the seedings were poorly done.

 

[Plebe]

I count 8 games in the first round in which a lower seed beat a higher seed. In the second round, I count 5 games with upsets, out of a total of sixteen games.

If you take out the top seeds, that means that in the second round nearly half (5 out of 12 games) resulted in upsets. In the first round, it meant that, if you take out the first three seeds, since they were clearly paired with inferior opponents, 8 of 20 games resulted in upsets. For both rounds, the seedings meant that the results could have been virtually replicated by throwing darts a a board with the teams' names on it.

So it seems that the NCAA seeding committee has hardly done a job that suggests competence. Would it be better to seed just the first four against the last four? And then just draw names out of a hat for all other games? I just can't believe that so many "upsets" are occurring. Seems more likely that the seedings were poorly done.

No. Occasional or even frequent upsets do not constitute proof of faulty seeding. Lower-seeded teams always win a certain percentage of the time. That percentage is typically higher on the men's side than the women's side, but in neither case does it mean that the committee botched the seeding. No one was ever going to argue before the tournament that Washington should be a #2 seed or that Maryland should be a #7 seed.

 

[UcMiami]

I count 8 games in the first round in which a lower seed beat a higher seed. In the second round, I count 5 games with upsets, out of a total of sixteen games.

If you take out the top seeds, that means that in the second round nearly half (5 out of 12 games) resulted in upsets. In the first round, it meant that, if you take out the first three seeds, since they were clearly paired with inferior opponents, 8 of 20 games resulted in upsets. For both rounds, the seedings meant that the results could have been virtually replicated by throwing darts a a board with the teams' names on it.

So it seems that the NCAA seeding committee has hardly done a job that suggests competence. Would it be better to seed just the first four against the last four? And then just draw names out of a hat for all other games? I just can't believe that so many "upsets" are occurring. Seems more likely that the seedings were poorly done.

Or in fact we have reach a level of parity that is what everyone has been clamoring about for the field as a whole.
If you look at the back end of the at large bids and throw away RPI as an evaluation criteria and just use results and actual visual evaluation and realize that beyond the #1 seeds teams are really inconsistent - you could have include about 20 teams in the discussion for the last 10 at large bids. Because they rely heavily on the RPI the committee had about ten teams in discussion for the last three or four spots. And if you look at the WNIT results you can see the same thing playing out with most of those 'bubble teams' that were not included with no regard for RPI and generally against the home court advantage. (Iowa was one of the committees near misses and lost at home to Ball State which was not in their discussion for example.)

For the rest of the seeds from 9 down to 2 - ASU was upset and was out of #2 consideration but then OSU, Stanford, and Louisville were also upset and they snuck back in to a 2 seed. Move back to the four seeds and you had the same thing with DePaul, FSU, and TAMU all playing themselves out of a four seed by being upset. Go further down the list and it just gets harder and harder to evaluate in any way. TN based on good wins should have been a three or four seed based on bad losses they should have been an eight or nine seed.

We can all point to upsets and close wins and come up with after the fact problems, but they would be different ones than we discussed when the brackets were announced generally. ASU always looked like a weak #2 but people were saying TN was a weak #7 too and no one would have dropped ASU down beyond a three or at worst a four, so the TN win would still have been an upset. BYU over-seeded, maybe, but Missouri was bubbly in most peoples evaluation - move BYU down a line and Missouri up a line and they could have met as 8/9 seeds or 9/8 seeds with the same result.

ND had one of the toughest 1 v 8/9 games I can remember in recent years and no one is arguing that Indiana was better than their seed. UCLA and OSU and Stanford all struggled against teams that most people think were pretty well seeded. Did anyone really complain about either Florida's or Albany's seeding or Maryland and Washington, or South Dakota St and Miami. Those were the real types of upsets that people have been craving for teams that were pretty well seeded - change the seed lines for those teams by one each to bring them closer and they are still significant upsets - and those seeds were really never in question until the results came up.

 

[Fairfield Fan]

No. Occasional or even frequent upsets do not constitute proof of faulty seeding. Lower-seeded teams always win a certain percentage of the time. That percentage is typically higher on the men's side than the women's side, but in neither case does it mean that the committee botched the seeding. No one was ever going to argue before the tournament that Washington should be a #2 seed or that Maryland should be a #7 seed.

I will have to respectfully differ. If there are "frequent upsets," then the upsets are not upsets. If a committee has properly judged which teams are better or worse than other teams, then most of the time the higher seeded team should win. But in this case, excepting the games in which the clearly superior teams were matched against clearly weaker teams, such as a 2 seed against a 14 seed, or a 3 seed against a 13 seed, the rate at which a lower seed defeated a higher seed is virtually half of the time. So it is demonstrably clear that the seedings are either wrong, or else there is such parity that the entire exercise of "seeding" is a fool's errand, because the differences between teams are so slight as to mean that it's simply not worth the effort.

So seed only the very top with the clearly weakest teams, and draw lots for the rest? Because common sense suggests that when a lower seed beats a higher seed half the time, something is not right.

 

[Plebe]

I will have to respectfully differ. If there are "frequent upsets," then the upsets are not upsets. If a committee has properly judged which teams are better or worse than other teams, then most of the time the higher seeded team should win. But in this case, excepting the games in which the clearly superior teams were matched against clearly weaker teams, such as a 2 seed against a 14 seed, or a 3 seed against a 13 seed, the rate at which a lower seed defeated a higher seed is virtually half of the time. So it is demonstrably clear that the seedings are either wrong, or else there is such parity that the entire exercise of "seeding" is a fool's errand, because the differences between teams are so slight as to mean that it's simply not worth the effort.

So seed only the very top with the clearly weakest teams, and draw lots for the rest? Because common sense suggests that when a lower seed beats a higher seed half the time, something is not right.

You are skewing your scientific sample by excluding only certain matchups involving the subjective concepts of "clearly superior" vs. "clearly weaker" teams. So let's examine this year's balance of upsets vs. chalk more closely:

In the first round:

• All 16 top-4 seeds won their opening-round games.
• Two 5-seeds lost to 12-seeds.
  
• All four of the 6-seeds won over the 11-seeds.
• Two 7-seeds lost to 10-seeds.
• All four of the 8-seeds lost to the 9-seeds.

This means that the higher seeds won 24 of the 32 first-round games. Four of the eight "upsets" involved a 9-seed beating an 8-seed, which most people would agree is not a significant upset at all. (And in any event, the distinction between being seeded 8 and 9 is entirely inconsequential.)

In the second round, 11 of the 16 games were won by the higher seed. Two of those five upsets, however, had extenuating circumstances: #4 Michigan State had to play at #5 Mississippi State, and #4 Texas A&M was without its second-leading scorer against #5 Florida State. Under those circumstances, we can hardly consider those two results to be surprising. That only leaves three true upsets in the second round: DePaul over Louisville, Tennessee over Arizona State, and Washington over Maryland.

I'm sure that if we were to compare these results to previous years' tournaments, the proportion of "upsets" to "chalk" would be very similar.

 

[UcMiami]

I will have to respectfully differ. If there are "frequent upsets," then the upsets are not upsets. If a committee has properly judged which teams are better or worse than other teams, then most of the time the higher seeded team should win. But in this case, excepting the games in which the clearly superior teams were matched against clearly weaker teams, such as a 2 seed against a 14 seed, or a 3 seed against a 13 seed, the rate at which a lower seed defeated a higher seed is virtually half of the time. So it is demonstrably clear that the seedings are either wrong, or else there is such parity that the entire exercise of "seeding" is a fool's errand, because the differences between teams are so slight as to mean that it's simply not worth the effort.

So seed only the very top with the clearly weakest teams, and draw lots for the rest? Because common sense suggests that when a lower seed beats a higher seed half the time, something is not right.

So what you are saying is the women's committee is pretty random this year but the men's committee has been a disaster for about the last twenty years?!
As the competitive nature of contests increases, the natural variation in performance for two relatively closely matched teams overlaps more and more - with the chance of the weaker team winning getting greater. In a series of single games between multiple teams where the stronger teams have a 60% win probability the weaker teams will end up winning 40% of the contests.

 

[Fairfield Fan]

So what you are saying is the women's committee is pretty random this year but the men's committee has been a disaster for about the last twenty years?!
As the competitive nature of contests increases, the natural variation in performance for two relatively closely matched teams overlaps more and more - with the chance of the weaker team winning getting greater. In a series of single games between multiple teams where the stronger teams have a 60% win probability the weaker teams will end up winning 40% of the contests.

I think the men's game is completely different due to the fact that the teams break up so often with top talent going pro. So less talented teams still are usually older, and have played together longer, while teams with better talent in power leagues are far less experienced, and haven't played together for nearly as long. So in that situation, upsets can be expected to occur with far more frequency. In the women's game, teams play together much longer, with talented teams retaining their players far longer. So there is a good deal less turnover, and considerably more predictability.

 

[Fairfield Fan]

You are skewing your scientific sample by excluding only certain matchups involving the subjective concepts of "clearly superior" vs. "clearly weaker" teams. So let's examine this year's balance of upsets vs. chalk more closely:

In the first round:

• All 16 top-4 seeds won their opening-round games.
• Two 5-seeds lost to 12-seeds.
  
• All four of the 6-seeds won over the 11-seeds.
• Two 7-seeds lost to 10-seeds.
• All four of the 8-seeds lost to the 9-seeds.

This means that the higher seeds won 24 of the 32 first-round games. Four of the eight "upsets" involved a 9-seed beating an 8-seed, which most people would agree is not a significant upset at all. (And in any event, the distinction between being seeded 8 and 9 is entirely inconsequential.)

In the second round, 11 of the 16 games were won by the higher seed. Two of those five upsets, however, had extenuating circumstances: #4 Michigan State had to play at #5 Mississippi State, and #4 Texas A&M was without its second-leading scorer against #5 Florida State. Under those circumstances, we can hardly consider those two results to be surprising. That only leaves three true upsets in the second round: DePaul over Louisville, Tennessee over Arizona State, and Washington over Maryland.

I'm sure that if we were to compare these results to previous years' tournaments, the proportion of "upsets" to "chalk" would be very similar.

So six of the eight teams seeded 7 or 8 lost to lower seeded teams. Sort of suggests that the committee really blew it. Or else, the metrics they were using were severely flawed. And for a third of teams seeded 5 and six lose to what are supposed to be teams that are considerably weaker?

Something's wrong. And if it is the same with previous years, then it only goes to prove that the seeding methodology is bunk. So many upsets means they're not upsets at all.

 

[Plebe]

So six of the eight teams seeded 7 or 8 lost to lower seeded teams. Sort of suggests that the committee really blew it. Or else, the metrics they were using were severely flawed. And for a third of teams seeded 5 and six lose to what are supposed to be teams that are considerably weaker?

Something's wrong. And if it is the same with previous years, then it only goes to prove that the seeding methodology is bunk. So many upsets means they're not upsets at all.

Once again, you're cherry-picking only those data that seem to suit your argument. What about the fact that all four six-seeds won? What about the fact that the 8's and 9's are virtually indistinguishable?

Imagine a 6-sided die with 4 sides painted red and 2 sides painted blue. One would expect, on average, the die to turn up "red" twice as many times as "blue." Someone then picks up the die and rolls it 6 times and, lo and behold, the die rolls "blue" 4 times and "red" only twice. Does this mean that the die-thrower used an incorrect throwing technique? You seem to be arguing that he must have.

 

[Vowelguy]

You have now posted this theory 3 times, and each time BY posters have told you that your reasoning is flawed. Since you don't want to listen to anyone here, I would suggest that you read a book (or take a class) on probability and statistics; you will learn that the results are not as rare as you think.

 

[Icebear]

Behind UCONN parity has grown greatly.

 

[Fairfield Fan]

Once again, you're cherry-picking only those data that seem to suit your argument. What about the fact that all four six-seeds won? What about the fact that the 8's and 9's are virtually indistinguishable?

Imagine a 6-sided die with 4 sides painted red and 2 sides painted blue. One would expect, on average, the die to turn up "red" twice as many times as "blue." Someone then picks up the die and rolls it 6 times and, lo and behold, the die rolls "blue" 4 times and "red" only twice. Does this mean that the die-thrower used an incorrect throwing technique? You seem to be arguing that he must have.

Your argument about throwing dice is precisely the point. This is random, while seedings are supposed to result in the higher seeds beating the lower seeds. I am not surprised that this happens with regularity. But it should be rare, not the rule.

And as you point out, if 8 and 9 seeds are virtually indistinguishable, then why go through the process? My point is that if for many of these seedings, we can agree that there's not much difference, then just seed the top teams and the bottom teams, and discard the fallacy that this is a detailed, data-driven process that gets it right most of the time.

Again, statistics show that there are too many "upsets" than a seeding process should be coming up with.

 

[Plebe]

Your argument about throwing dice is precisely the point. This is random, while seedings are supposed to result in the higher seeds beating the lower seeds. I am not surprised that this happens with regularity. But it should be rare, not the rule.

Seedings are not intended to ensure that one team will win over another. They are merely intended to distribute the quality of competition evenly across a bracket, and to ensure that the best teams don't have to face each other in the earliest rounds. Go look at any tennis tournament, and count how many seeded players lose to lower-seeded or unseeded players. It happens all the time. It doesn't mean they were mis-seeded. It just means that athletic contests are, to a certain degree, inherently unpredictable.

And as you point out, if 8 and 9 seeds are virtually indistinguishable, then why go through the process? My point is that if for many of these seedings, we can agree that there's not much difference, then just seed the top teams and the bottom teams, and discard the fallacy that this is a detailed, data-driven process that gets it right most of the time.

Again, statistics show that there are too many "upsets" than a seeding process should be coming up with.

I'm sorry, but it does get it right most of the time: 24-of-32 and 11-of-16 is, by definition, most of the time.

Most people complain that the women's tournament doesn't have enough upsets. I find it exceedingly ironic that someone is now finding cause to complain that there are too many upsets.

 

[RSHERMVIKES]

I finally watched the Washington 75 / MD 64 game and besides joy at Brenda (speech maker) Frese losing, I noticed that anytime in crunch time her asst's one tall blond lady and a guy did most of the talking, adding to other info, and basically controlling the huddles!
Remember in 2006 when MD won their nat'l title the real HEAD COACH was Jeff Walz now of L'ville fame he ran most timeouts and gave valuable info to the players.
Brenda gets all the credit and gives all the speeches and is just a figure head while the real brains of the outfit do the dirty work!
Everytime she's on the screen I hold my wallet in both my hands as she's so sleezy, I'm afraid she'll pick my pocket through the TV screen.
I told this story before but bares repeating- - - Brenda got the MN head coaching job (Lindsay Whalen's junior year) and at the introduce presser said, "The MN job was her DREAM JOB and she wanted to lead MN into the NCAA's for years to come." After that 1st year she took the MD job and at that introduce presser she started off "This MD job is her DREAM come true job, and she plans on leading MD into many NCAA tournaments in the future!"
Brenda was rumored to have stolen Harper and Langhorne from UCONN/Geno for MD at basically the last moment before signing day. Geno said something about coaches who call athletes and make up stories to get the players to go to their school! Never named names but the press inferred Brenda Frese!

 

[meyers7]

And as you point out, if 8 and 9 seeds are virtually indistinguishable, then why go through the process?

Ummm because when you go from 1 to 16, you just can't skip 8 and 9. They are pretty much indistinguishable because the are at the middle. If you were seeding 1-8 then 4/5 would be as close as you get. If you were seeding 1-32, then 15/16 would be the closest. It's just math.

My point is that if for many of these seedings, we can agree that there's not much difference, then just seed the top teams and the bottom teams, and discard the fallacy that this is a detailed, data-driven process that gets it right most of the time.

But it does get it right most of the time. Not sure why you can't grasp this?

 

[Kibitzer]

Reading the arguments in this thread make my head hurt.

 

[Vowelguy]

Geez I thought the myth that Walz did everything and Frese was nothing without him was finally dead and gone.

I would love to hear the explanation then of how 4th seeded Maryland beat the 3rd seeded Louisville team returning all its players from the runnerup team and was playing on ITS HOME FLOOR in 2014. For example.

 

[Wally East]

Seedings are based on past performance and do not predict future results. It's really that simple. After that, teams will overperform and underperform because humans are involved. That's also pretty simple.

 

[Fairfield Fan]

You may argue about seedings, and whether the NCAA needs to use other metrics, or only use those metrics to seed the top few and bottom few. But the title of the thread was right on target!

Wonder if this tournament has included the most upsets, and big upsets, in its history. Anyone willing to take a stab at it?