NCAA Tournament: 2011 Format

Anyone have any thoughts on the 1+2+1 format? Did it help strong teams? Did it hurt them? Looking at the bracket, Santa Clara bounced, OSU bounced, Pepperdine bounced as a result of having tough first round matches due to geography. Looks like the west coast may have suffered a bit.

I haven't really checked rigorously, but it looks like the West might have gotten slightly harder brackets just because there aren't so many teams out here and travel is still an issue, but I don't think it is a big deal. you are going to have to beat somebody to advance anyway.


The ACC is having to travel a lot just because there are so many teams, so i think they are getting screwed a little on travel.

But I think next weekend will be really cool. There looks like there are some pretty neat brackets to sort themselves out. it should make for an exciting weekend. the four team pods are of a higher quality than the old system

Geezaldinho wrote:I haven't really checked rigorously

That is so unlike you!

Looking at the first round games with at 64 team draw you expect a 32 seed to play a 33 if you could make a perfect bracket so the total of the two teams seeds should come in 65. We don't know the exact seeding but if you use RPI as a proxy, then the toughest 1st round games were WV 17 rpi vs. VT 30 (47 total), A&M 3 seed rpi 11, vs. LSU 33 rpi (44 total), UNC 3 seed 12 rpi vs. W&M 34 (46 total), LB St 31 vs. Pep 3 seed 10 rpi (41 total). Notice #3 seeds had strong opponents.

Second round games: we have Ill 19 rpi vs. OK St 7 (26 total compared to 33 in perfect bracket). UNC 12 vs. Baylor 15 (27 total). UNC not getting the patsies for a change.

You have 9 games this round involving teams that "upset" a higher rpi team. Those teams are CAL, WSU, VT, Lou, UP, SD, LB St, OH St, and Geo. Those teams playing the upset winners don't have as tough a match up on paper. Tells you something about the value of rpi when 9 of 32 games won by the poorer rpi team particularly when higher seed hosted (Only exception I believe was Lou which hosted Dayton.)

gnarly wrote: Tells you something about the value of rpi when 9 of 32 games won by the poorer rpi team

Of the 9 "upsets," 4 were ties, and one other, as you point out, was won by the home team. In a sport that statisticians rate as one of the more unpredictable, I think you would average that many upsets in any field of 64, even if they're perfectly seeded.

dwm wrote:

gnarly wrote: Tells you something about the value of rpi when 9 of 32 games won by the poorer rpi team

Of the 9 "upsets," 4 were ties, and one other, as you point out, was won by the home team. In a sport that statisticians rate as one of the more unpredictable, I think you would average that many upsets in any field of 64, even if they're perfectly seeded.

That actually doesn't help the argument.

Home field is actually pretty powerfully in women's soccer.

A couple of the ratings value home field at about .6 goals. Albyn Jones used to figure it at about 60-75 ratings points, where 100 pts is the difference between 1:1 odds and 2:1 odds.

That would swing a match between evenly rated opponents to about 60-65% in favor of the home team. With the home team having the higher rating, home team Odds will often be significantly higher. One could expect the higher rated home team in the first round of a perfectly seeded field to win at a rate in the high 80's or low 90's, even including odds for ties.

So 9/32 upset rate is actually pretty high. (Of course, geography had to do as much as ratings with the first round seeds.)

It's actually high enough to swing the odds of a potential matchup between #1 seed FSU and #4 seed Memphis to the home team.

As I brought up to the person on BS who first brought that up, the home team scenarios also reduce the odds of that matchup even happening.

Geezaldinho wrote:Home field is actually pretty powerfully in women's soccer.

I'm wondering, however, if there is a counterbalance to home field advantage when it comes to shootouts. As I think back on Pilot women's shootouts that I can remember, they have a "perfect" record:

• Lost every shootout at home (2003 Santa Clara, 2010 Washington).
  
• Won every shootout on the road or at a neutral site (2002 Stanford, 2005 Penn State, 2011 Oregon State).

In that context, I find it interesting that last weekend all four shootouts were won by the visiting team (USD@UCI, Cal@SCU, WSU@Kentucky, UP@OSU).

Is there something about the pressure of a shootout in front of a team's own fans?

That is interesting data point and something to pay attention to in the future. I know that going first is preferred in shootouts. eProf, you may be the author of a new theory.

Agree with the comment that 9 upsets in 32 games seems pretty high. While I would like to think the NCAA used RPI, I'm not sure that geo didn't get more weight. It will be interesting to see interesting to see how the next two round go. My suspicion is the final eight will contain most of the top seeds which is what you want. Unfortunately, a number of quality teams probably got bounced earlier than they would have with the previous format.

the seeds (top 16) meet each other in the same round they used to. The major difference I see is that the seeds will meet on the second game of a weekend.

How they get trough the first game will affect that game a lot, and the high seeds THEORETICALLY have the easier game on Friday.

So I expect fewer upsets on Sunday. One fly in the ointment is that ACC teams aren't all that used to playing with only one day's rest. they schedule Thurs-Sun during the conference season.

UCLA lost their shootout at home. they are the first #1 to drop out.

the Cal _BC shootout was technically Neutral. maybe the theory extends to distance from home?

7 seeded teams are gone. 9 remain.

Of the possible 8 matchups between teams on Sunday, only 2 contests will have both seeded teams playing.

The average number of seeded teams left at this stage since they instituted the 64 team format is 11.4.

Actually, UCLA was a #2. The #1s are Stanford, Duke, Wake Forest, and Florida State.

UPSoccerFanatic wrote:Actually, UCLA was a #2. The #1s are Stanford, Duke, Wake Forest, and Florida State.

Ha!

got their seed and their score confused on the interactive bracket.

Geez, that happens to all of us. It just doesn't happen to you.

Geez made a mistake??!? My world is falling apart. Now I'm wondering - Is UPSF a mathematical illiterate, is Soccer Joy really a closet UCLA fan? I have to go meditate now.

Quite the contrary. We were overjoyed to see the much hated UCLA got knocked out by a WCC team!!!

The world is almost back on its axis

The only difference is that when we were winning, nobody noticed I made mistakes. We have a sub-par season and you guys have turned into petty, mean spirited nit- pickers.


Yeah, that's it.

You made OTHER mistakes BEFORE this season?? And, nobody noticed??

I guess you're right - nobody did. But, we do know about them now that you mentioned it, and I think I am going to go back through all your past posts and start retroactively nit picking. All 6657 posts - - - well, maybe not.

fan from afar wrote:You made OTHER mistakes BEFORE this season?? And, nobody noticed??

I guess you're right - nobody did. But, we do know about them now that you mentioned it, and I think I am going to go back through all your past posts and start retroactively nit picking. All 6657 posts - - - well, maybe not.

Good lord! I wouldn't even do that.


I can pretty much guarantee you that the end result will be more painful for you than for me...

Maybe you could give us a list of your top 10 mistakes just to save us the trouble of looking for them.

Mistakes?

What mistakes?

Well, the Geez may have made one mistake, and I have made plenty of them with my RPI stuff, but none of them compares to the mistake the NCAA made related to D1 women's soccer ratings last year and this.

PS - This is a teaser! More to come.

The following attendance numbers for this weekend's games suggest to me that maybe the new format is not such a good idea. The first table is for games in which a host team played. The second table is for games in which no host team played, so that the game was at a neutral site.

In the neutral site attendance table, for some of the games I've listed the attendance as 0. This is because the host school did not separately track game attendance for the two games on Friday night, but instead reported the same attendance figure for both games. When that happened, I attributed the entire attendance to the host team's game and listed the neutral site game's attendance as 0. Still, if you look at the numbers where there was a separate attendance count for a neutral site game, you'll get the idea.

Sunday, November 20, 2011 Home Attendance

Attendance/Home Team/Away Team

Code:

1524   Stanford v BostonCollege
1094   VirginiaU v VirginiaTech
958    OklahomaState v MarylandU
949    Duke v OhioState
588    WakeForest v PennState

Friday and Sunday, November 18 and 20, 2011 Neutral Site Attendance

Attendance/Team/Team

Code:


975   CaliforniaU v BostonColege
240   VirginiaTech v TexasA&M
164   Marquette v PennState
111   MarylandU v Auburn
0     PortlandU v FloridaState
0     NorthCarolinaU v Baylor
0     MiamiFL v LongBeachState
0     OhioState v Milwaukee
598   LongBeachState v SanDiegoU
245   UCF v NorthCarolinaU

20-Nov Sunday 101 FloridaState Louisville

I heard that the Geez thought he made a mistake once, but he turned out to be mistaken.

No, it turns out, I was wrong.