Discussion thread: Which is the hardest bracket?

[brevity]

I'm not a big fan of the Eye Test -- it seems like an excuse for Digger Phelps to disregard all the stats and focus instead on name recognition of a team, coach, or player. But some variation of the Eye Test might be helpful to weigh the toughness of the regions.

Namely: which team of a particular seed number is the toughest? Which is the weakest? Go from 1-8 or so, and then figure out if one of the regions tends to have more than its fair share of the teams you chose as the toughest.

For example, we can use Pomeroy numbers to list each region from toughest to weakest by numerical seed.

1: East (Ohio State, 1); West (Duke, 2); Southwest (Kansas, 3); Southeast (Pittsburgh, 5)
8: Southwest (UNLV, 22); East (George Mason, 26); West (Michigan, 40); Southeast (Butler, 54)
4: West (Texas, 4); East (Kentucky, 7); Southeast (Wisconsin, 9); Southwest (Louisville, 12)
5: East (West Virginia, 21); West (Arizona, 25); Southeast (Kansas State, 29); Southwest (Vanderbilt, 32)

3: Southwest (Purdue, 9); East (Syracuse, 11); Southeast (BYU, 13); West (Connecticut, 17)
6: West (Cincinnati, 23); Southwest (Georgetown, 31); Southeast (St. John's, 35); East (Xavier, 37)
2: West (SDSU, 6); Southwest (Notre Dame, 10); East (UNC, 14); Southeast (Florida, 19)
7: East (Washington, 15); West (Temple, 38); Southwest (Texas A&M, 45); Southeast(UCLA, 53)

Some interesting trends emerge. The East region, with the exception of Xavier and UNC, has the toughest or second toughest team by seed number. Meanwhile, the Southeast region has the weakest or second weakest team for every seed number, with no exception. The West and the Southwest are fairly close: the 1/8/4/5 subregion looks tougher in the West, while the 3/6/2/7 subregion looks tougher in the Southwest.

The numbers here seem to reinforce what most of us have been thinking, so maybe there's something to the Eye Test. And, let's face it, that's basically what we use when we fill out our brackets.

[A-Tex Devil]

I don't disagree with your general point, but Texas going 4-4 in their last 8 games. You have to go back nineteen games to find Florida's last 4 losses. Recent play does have a big effect on seeding.

By the way, Kenpom doesn't take recent play into account does he? i.e. would a four losses in December be weighted the same as a team losing its last four games?

Fair enough. Texas ended the season about the same way, oh, that Pittsburgh did. Florida hasn't sniffed the top 8 in any measurable index all year that I am aware of. They just made the top 10 in the RPI. They are probably a 3, and maybe Texas is too.

The committee seems to glom on to one factor when misseeding a team, whether it be UCONN (who even with the Big East tourney win is overseeded considering it finished in the BOTTOM THIRD of its conference), Texas, Florida, or the quartet of Lukcy Batsards that had no business being in the tourney - VCU, Georgia, USC and UAB.

Kenpom, I believe, does have a recency bias in his calculations and weights more recent games more heavily. I could be wrong, but I would think he'd certainly include this considering how thorough he is.

[Delaware]

A bit of a Geek method, but I have done this the past few years.

I take Kenpom, Sagarin and RPI and average them. The result is my own personal S-Curve which looks like this:

1 Ohio St (E)
2 Kansas (SW)
3 Duke (W)
4 San Diego St (W)
5 Pitt (SE)
6 Texas (W)
7 BYU (SE)
8 Kentucky (E)
9 Notre Dame (SW)
10 Purdue (SW)
11 UNC (E)
12 Wisconsin (SE)
13 Syracuse (E)
14 Louisville (SW)
15 Florida (SE)
16 Conn (W)

This year these teams happen to make up the full first 4 seed lines.

If you sum up the values for each region you get:
West = 29
East = 33
SW = 35
SE = 39

[FerryFor50]

Fair enough. Texas ended the season about the same way, oh, that Pittsburgh did. Florida hasn't sniffed the top 8 in any measurable index all year that I am aware of. They just made the top 10 in the RPI. They are probably a 3, and maybe Texas is too.

The committee seems to glom on to one factor when misseeding a team, whether it be UCONN (who even with the Big East tourney win is overseeded considering it finished in the BOTTOM THIRD of its conference), Texas, Florida, or the quartet of Lukcy Batsards that had no business being in the tourney - VCU, Georgia, USC and UAB.

Kenpom, I believe, does have a recency bias in his calculations and weights more recent games more heavily. I could be wrong, but I would think he'd certainly include this considering how thorough he is.

Totally agree with this.

It's baffling how the seedings were formulated. Florida gets a 2 despite losing to Kentucky in the SEC final (handily, I might add) and Kentucky gets a 4?

BYU ends the season with 2 blowout losses and loses their starting center. But they get a 3?

I'd also question UNC's #2 seed... who have they beaten, really, outside of Duke and a narrow win early in the year against Kentucky? They struggled all tournament and have struggled all year against some lesser competition.

It's like the committee designed the brackets to be more upset-friendly since last year's bracket went mostly by design...

[Troublemaker]

Using only my own subjective opinion (eye test), here are the Final Four paths for each #1 seed if no upsets occur.

KU: Illinois -----> Louisville -----> Notre Dame
Duke: Mich/Tenn -----> Texas -----> SDSU
tOSU: Vill/Mason -----> Kentucky -----> UNC
Pitt: Butler -----> Belmont -----> Florida

I have listed the 4 paths from toughest to easiest in my subjective opinion.

KU will be tested thoroughly at all 3 levels.
Duke has very tough Sweet 16 and Elite 8 opponents, so 2 levels of harshness.
tOSU has a very talented Sweet 16 opponent but a weak 2 seed, so 1 level of harshness.
Pitt seemingly just has an easy path.

I would trade Duke's bracket for tOSU's or Pitt's but not Kansas'

[FerryFor50]

Using only my own subjective opinion (eye test), here are the Final Four paths for each #1 seed if no upsets occur.

KU: Illinois -----> Louisville -----> Notre Dame
Duke: Mich/Tenn -----> Texas -----> SDSU
tOSU: Vill/Mason -----> Kentucky -----> UNC
Pitt: Butler -----> Belmont -----> Florida

I have listed the 4 paths from toughest to easiest in my subjective opinion.

KU will be tested thoroughly at all 3 levels.
Duke has very tough Sweet 16 and Elite 8 opponents, so 2 levels of harshness.
tOSU has a very talented Sweet 16 opponent but a weak 2 seed, so 1 level of harshness.
Pitt seemingly just has an easy path.

I would trade Duke's bracket for tOSU's or Pitt's but not Kansas'

Belmont?

You know something we don't?

[Troublemaker]

Belmont?

You know something we don't?

Haha, yeah, that subregion is going to be very interesting.

I'd probably rank the teams Belmont > Utah St > Wiscy > KSU

But they're all very close in quality and any of them could emerge from that subregion. And not one of them is as threatening to a 1 seed as Texas, UK, and Louisville would be as Sweet 16 opponents, imo. It's been said a bunch of times now but Pitt really did seem to get a break this year.

[FerryFor50]

Haha, yeah, that subregion is going to be very interesting.

I'd probably rank the teams Belmont > Utah St > Wiscy > KSU

But they're all very close in quality and any of them could emerge from that subregion. And not one of them is as threatening to a 1 seed as Texas, UK, and Louisville are, imo.

Interesting... I guess you see Belmont as maybe this year's Butler?

I think Utah St might be a little better than Belmont. And speaking of Butler, I see them as better than Belmont, despite their slow start. I think they're finally starting to gel a bit.

[CDu]

I think Pitt clearly got the easiest draw. They have the weakest combination of 2, 3, and 4 seeds. I'd argue they have the second weakest 2 seed (to our #2), the weakest #3 and weakest #4. Their #5 has looked tougher recently, but not tougher than the #4 seeds (the #1 seed would only play at most one of the 4/5 seeds).

Our first weekend isn't unreasonable. It won't be easy (nothing is easy in the tournament), but it's manageable. The second weekend could be very tough. I think we got the toughest of the possible #4 seeds. If we get by them, I don't think our 2/3 combination is any harder (perhaps easier) than that of KU or OSU.

But, it's just a matter of playing as well as we can each game. Coach K is great at getting his team to compartmentalize each phase of the process. And several of the guys have experience in each phase of the tournament.

[ACCBBallFan]

I do the same thing but I use 3 factors, the other being Pomeroy all at same weight. Same top 16 in slightly different order that I will append to yours. Top 5 and bottome 2 exactly same

The one with largest swing was UK 8 vs. 11, E verybody else +/- 1 or same.

A bit of a Geek method, but I have done this the past few years.

I take Kenpom, Sagarin and RPI and average them. The result is my own personal S-Curve which looks like this:

1 Ohio St (E) -1
2 Kansas (SW) -2
3 Duke (W) -3
4 San Diego St (W) -4
5 Pitt (SE) -5
6 Texas (W) -7
7 BYU (SE) -6
8 Kentucky (E)-119 Notre Dame (SW) -8
10 Purdue (SW) -9
11 UNC (E) -10
12 Wisconsin (SE) -12
13 Syracuse (E) -14
14 Louisville (SW) -13
15 Florida (SE) -15
16 Conn (W) -16

This year these teams happen to make up the full first 4 seed lines.

If you sum up the values for each region you get:
West = 29
East = 33
SW = 35
SE = 39

[DallasDevil]

This just came up on a chat with Joe Lunardi:

"Jorell (Hartsville,SC) [via mobile]


Today the tournament committee chair said that Wright returning for Georgetown and Singleton returning for FSU influenced their teams seeding. If Coach K would have said that Kyrie Irving would return for the tournament do you think that Duke could have possibly jumped Pitt for the third number 1 seed?

Joe Lunardi (4:06 PM)


That's a distinction without a difference, Jorell. If I'm Duke, I'm perfectly happy to open in Charlotte and play in what appears to be a softer West Region."

The West is softer than the Southeast? I almost get the feeling that the committee could have put Ohio State, Notre Dame, and Texas in our region and everyone would still complain about what an easy path we have to the Final Four.

[superdave]

Kenpom has updated his rankings and has listed seeds next to team names, and Texas sticks out like a sore thumb as a 4 seed when you scan the page. That team is quite underseeded.

It'll be a tough road. SDSU is probably underestimated nationally the same way Butler was last year. Remember, Butler was 33-4 with a 25 game winning streak going into the championship game against Duke. Some analysts were predicting an easy win for Duke, but when a team like Butler wins games with that kind of regularity, they almost certainly have built up amazing chemistry and institutional excellence that makes their whole better than the sum of their parts. Well, this season SDSU is a higher-ranked Kenpom team than Butler of last season and would be an incredible 35-2 if they make it to the Elite 8. I'm definitely wary of them.

But, first things first. Hampton, then Michigan/Tennessee.

I could see Duke having to really play flawless ball, and make some in-game adjustments, to beat SDSU. Duke can do it though. BYU out-executed (and shot well) against them twice. But Duke would have to control the tempo and not let it turn into a track meet. Control the boards, make them play tough D.

[uh_no]

http://fivethirtyeight.blogs.nytimes...luck-is-vital/

Take a look down at the teams with the best and worst draws...i don't exactly know where the numbers come from, but 538 is well known for their statistics work...

interesting that 2 of the 1 seeds have some of the worst draws while the other 2 have some of the best in the tournament!

also of note is that UNC is up their for one of the easiest draws