collapse

* Recent Posts

2024 NCAA Tournament Thread by Plaque Lives Matter!
[Today at 01:02:54 AM]


45 minutes ago at the Dallas Westin by MuggsyB
[Today at 12:19:24 AM]


2024 Coaching Carousel by Plaque Lives Matter!
[Today at 12:10:57 AM]


2024 Transfer Portal by CountryRoads
[Today at 12:05:42 AM]


Are we still recruiting anyone for the 24-25 season. by Don_Kojis
[Today at 12:04:21 AM]


Where is Marquette? by marqfan22
[March 28, 2024, 09:29:52 PM]


Chicago bars for Fri game by Daniel
[March 28, 2024, 08:47:22 PM]

Please Register - It's FREE!

The absolute only thing required for this FREE registration is a valid e-mail address.  We keep all your information confidential and will NEVER give or sell it to anyone else.
Login to get rid of this box (and ads) , or register NOW!


Author Topic: NCAAT Expected Wins by seed -- who's over/underperforming?  (Read 927 times)

mikekinsellaMVP

  • Registered User
  • Starter
  • ***
  • Posts: 232
NCAAT Expected Wins by seed -- who's over/underperforming?
« on: March 21, 2017, 11:07:52 PM »
I saw some discussion in the "Fire K" thread about overperforming/underperforming relative to seed.  I've never liked the "losing to a lower seed means you underperformed" metric, mostly because winning several games in a single elimination tournament (as high seeds are expected to do) is really, really hard.  Just look at FiveThirtyEight's pre-tournament odds -- Northwestern had the worst odds of any 5-8 seed "matching their seed" (45% chance of winning their first round game), while Gonzaga had the best odds of any 1 seed doing the same (41% chance of reaching FF).

Channeling my inner Chicos, I figured there must be some way to weight seeds to account for the "crapshoot" factor.  Fortunately for me, some guys at U of I have already done the heavy lifting.  Aggregating every tourney result from 1985 (the first 64-team tourney) to 2016, they've come up with an Expected Wins value for each seed.  (It's pretty straightforward:  128 1-seeds have combined to win 429 tournament games, averaging 3.35 Expected Wins per 1-seed.  And so forth.)  Having too much time on my hands today, I looked at several coaches’ tournament records over the last ten years (2007-2016) and compared their seed to their performance each year to see who is performing above, at, and below expectations.  I've subjectively said +/-2 games is the magic number to over/underperform.  A few selected names:
 
-------------------------
 
Coach K (Duke)
Appearances: 10
Average seed: 2.4
Expected wins: 24.13
Actual wins: 22
O/U: -2.13
Performance:  Underachieving, barely.  This one is largely personal interpretation: are two titles sufficient to offset three first round exits as a two or three seed?  I’d say hell yes, but I also don’t have the expectations of the Duke fan base.
 
Bill Self (Kansas)
Appearances: 10
Average seed: 1.5
Expected wins: 29.1
Actual wins: 27
O/U: -2.1
Performance:  Underachieving, barely.  Very similar to K, but the same number of Final Fours and one less title despite better seeding (never lower than 3).  This may be an indictment of the competition in the Big 12 as much as it is about Self’s coaching.
 
Roy Williams (UNC)
Appearances: 9
Average seed: 2.78
Expected wins: 22.56
Actual wins: 28
O/U: +5.44
Performance:  Overachieving.  The Heels have made at least the Elite Eight every year they have been a 1 or 2 seed (6 times).  One less title and appearance than K, but 6 more wins.  Maybe he learned something about consistency from his athletes’ identical term papers.
 
Rick Pitino (Louisville)
Appearances: 9
Average seed: 4.00
Expected wins: 16.39
Actual wins: 22
O/U: +5.61
Performance:  Overachieving.  Of the guys who have had their current gig for 10+ years, he’s the highest above expected wins.  Party at Porcini’s?
 
John Calipari (multiple schools)
Appearances: 9 (3 Memphis, 6 Kentucky)
Average seed: 2.67
Expected wins: 22.03
Actual wins: 33
O/U: +10.97
Performance:  DAMN.  Love him or hate him, the guy’s postseason teams win basketball games.  Even if you take out his 2014 title appearance as an absurdly underseeded 8, he’s still +6.7.  Want to vacate Memphis’ 2008 title appearance, too?  Still +5.  Coach Cal’s teams are unreal.
 
Gregg Marshall (multiple schools)
Appearances: 6 (1 Winthrop, 5 Wichita State)
Average seed: 7.33
Expected wins: 7.07
Actual wins: 9
O/U: +1.93
Performance:  Almost overachieving, but with a catch.  Final Four runs as a high seed do wonders for your numbers; without the FF appearance as a 9 in 2013, Marshall is -1.51.
 
Shaka Smart (multiple schools)
Appearances: 6 (5 VCU, 1 Texas)
Average seed: 7.67
Expected wins: 5.34
Actual wins: 6
O/U: +0.66
Performance:  Gregg Marshall redux, but a step down.  Without the 2011 Final Four run as an 11 seed, Shaka is -2.76 in 5 appearances – not good.
 
Jay Wright (Villaova)
Appearances: 9
Average seed: 5.44
Expected wins: 14.55
Actual wins: 15
O/U: +0.45
Performance:  Meeting expectations.  Jay is obviously a great coach, but last year’s title appears to be an outlier when it comes to postseason success (-3.15 for 2007-2015).
 
John Thompson III (Georgetown)
Appearances: 7
Average seed: 3.14
Expected wins: 13.48
Actual wins: 7
O/U: -6.48
Performance:  Yuck.
 
Crean (multiple schools)
Appearances: 6 (2 Marquette, 4 Indiana)
Average seed: 5.67
Expected wins: 8.51
Actual wins: 7
O/U: -1.51
Performance:  Not terrible, but not great.  By comparison, Crean was -0.66 for his 5 appearances during his MU tenure.
 
Buzz (Marquette)
Appearances: 5
Average seed: 5.80
Expected wins: 6.44
Actual wins: 8
O/U: +1.56
Performance:  On the whole, Buzz was undoubtedly the better postseason coach of the previous two regimes. 
 
 
...and for a young coach who just made his first tournament...
 
Wojo (Marquette)
Appearances: 1
Average seed: 10
Expected wins: 0.64
Actual wins: 0
O/U: -0.64
Performance:  GET THE PITCHFORKS!!!
 

rocky_warrior

  • Global Moderator
  • All American
  • *****
  • Posts: 9002
Re: NCAAT Expected Wins by seed -- who's over/underperforming?
« Reply #1 on: March 21, 2017, 11:29:19 PM »
Actually, I'd say Wojo nailed that one.  Marquette was winning for about 64% of the game!

Benny B

  • All American
  • *****
  • Posts: 5969
Re: NCAAT Expected Wins by seed -- who's over/underperforming?
« Reply #2 on: March 22, 2017, 09:27:10 AM »
Actually, I'd say Wojo nailed that one.  Marquette was winning for about 64% of the game!

Wow, I'm very concerned for Benny.  Being able to mimic Myron Medcalf's writing so closely implies an oncoming case of dementia.

GoldenDieners32

  • All American
  • *****
  • Posts: 2053
Re: NCAAT Expected Wins by seed -- who's over/underperforming?
« Reply #3 on: March 22, 2017, 10:29:58 AM »
I saw some discussion in the "Fire K" thread about overperforming/underperforming relative to seed.  I've never liked the "losing to a lower seed means you underperformed" metric, mostly because winning several games in a single elimination tournament (as high seeds are expected to do) is really, really hard.  Just look at FiveThirtyEight's pre-tournament odds -- Northwestern had the worst odds of any 5-8 seed "matching their seed" (45% chance of winning their first round game), while Gonzaga had the best odds of any 1 seed doing the same (41% chance of reaching FF).

Channeling my inner Chicos, I figured there must be some way to weight seeds to account for the "crapshoot" factor.  Fortunately for me, some guys at U of I have already done the heavy lifting.  Aggregating every tourney result from 1985 (the first 64-team tourney) to 2016, they've come up with an Expected Wins value for each seed.  (It's pretty straightforward:  128 1-seeds have combined to win 429 tournament games, averaging 3.35 Expected Wins per 1-seed.  And so forth.)  Having too much time on my hands today, I looked at several coaches’ tournament records over the last ten years (2007-2016) and compared their seed to their performance each year to see who is performing above, at, and below expectations.  I've subjectively said +/-2 games is the magic number to over/underperform.  A few selected names:
 
-------------------------
 
Coach K (Duke)
Appearances: 10
Average seed: 2.4
Expected wins: 24.13
Actual wins: 22
O/U: -2.13
Performance:  Underachieving, barely.  This one is largely personal interpretation: are two titles sufficient to offset three first round exits as a two or three seed?  I’d say hell yes, but I also don’t have the expectations of the Duke fan base.
 
Bill Self (Kansas)
Appearances: 10
Average seed: 1.5
Expected wins: 29.1
Actual wins: 27
O/U: -2.1
Performance:  Underachieving, barely.  Very similar to K, but the same number of Final Fours and one less title despite better seeding (never lower than 3).  This may be an indictment of the competition in the Big 12 as much as it is about Self’s coaching.
 
Roy Williams (UNC)
Appearances: 9
Average seed: 2.78
Expected wins: 22.56
Actual wins: 28
O/U: +5.44
Performance:  Overachieving.  The Heels have made at least the Elite Eight every year they have been a 1 or 2 seed (6 times).  One less title and appearance than K, but 6 more wins.  Maybe he learned something about consistency from his athletes’ identical term papers.
 
Rick Pitino (Louisville)
Appearances: 9
Average seed: 4.00
Expected wins: 16.39
Actual wins: 22
O/U: +5.61
Performance:  Overachieving.  Of the guys who have had their current gig for 10+ years, he’s the highest above expected wins.  Party at Porcini’s?
 
John Calipari (multiple schools)
Appearances: 9 (3 Memphis, 6 Kentucky)
Average seed: 2.67
Expected wins: 22.03
Actual wins: 33
O/U: +10.97
Performance:  DAMN.  Love him or hate him, the guy’s postseason teams win basketball games.  Even if you take out his 2014 title appearance as an absurdly underseeded 8, he’s still +6.7.  Want to vacate Memphis’ 2008 title appearance, too?  Still +5.  Coach Cal’s teams are unreal.
 
Gregg Marshall (multiple schools)
Appearances: 6 (1 Winthrop, 5 Wichita State)
Average seed: 7.33
Expected wins: 7.07
Actual wins: 9
O/U: +1.93
Performance:  Almost overachieving, but with a catch.  Final Four runs as a high seed do wonders for your numbers; without the FF appearance as a 9 in 2013, Marshall is -1.51.
 
Shaka Smart (multiple schools)
Appearances: 6 (5 VCU, 1 Texas)
Average seed: 7.67
Expected wins: 5.34
Actual wins: 6
O/U: +0.66
Performance:  Gregg Marshall redux, but a step down.  Without the 2011 Final Four run as an 11 seed, Shaka is -2.76 in 5 appearances – not good.
 
Jay Wright (Villaova)
Appearances: 9
Average seed: 5.44
Expected wins: 14.55
Actual wins: 15
O/U: +0.45
Performance:  Meeting expectations.  Jay is obviously a great coach, but last year’s title appears to be an outlier when it comes to postseason success (-3.15 for 2007-2015).
 
John Thompson III (Georgetown)
Appearances: 7
Average seed: 3.14
Expected wins: 13.48
Actual wins: 7
O/U: -6.48
Performance:  Yuck.
 
Crean (multiple schools)
Appearances: 6 (2 Marquette, 4 Indiana)
Average seed: 5.67
Expected wins: 8.51
Actual wins: 7
O/U: -1.51
Performance:  Not terrible, but not great.  By comparison, Crean was -0.66 for his 5 appearances during his MU tenure.
 
Buzz (Marquette)
Appearances: 5
Average seed: 5.80
Expected wins: 6.44
Actual wins: 8
O/U: +1.56
Performance:  On the whole, Buzz was undoubtedly the better postseason coach of the previous two regimes. 
 
 
...and for a young coach who just made his first tournament...
 
Wojo (Marquette)
Appearances: 1
Average seed: 10
Expected wins: 0.64
Actual wins: 0
O/U: -0.64
Performance:  GET THE PITCHFORKS!!!

Fire Wojo