Interesting Tournament Upset Predictor | The Boneyard

Interesting Tournament Upset Predictor

Status
Not open for further replies.

OkaForPrez

Really Popular Poster
Joined
Aug 28, 2011
Messages
5,207
Reaction Score
26,712
Interesting stat insight in research today: I have to update my dataset, but from 2002 - 2011 top seeded teams (1-5) who lost in the first round had an average absolute difference of 40 in KenPom AdjO vs. AdjD rank. Top seeded teams who won in the first round have an avg difference of 20. Meaning the more unbalanced a team is on one side of the ball vs. the other, the more likely they are to get sniped.

Look at recent history (adjO, Adj D, Diff):

2014 Duke - (2, 116, 114) Ohio State - (128, 3, 125) VCU - (107, 6, 101)
2013 GTown - (78, 2, 76) Wisconsin ( 108, 1, 107)
2012 Duke - (10, 81, 71) Missouri (1, 146, 145)

Note these are post tourney ranks because I'm not a kenpom subscriber and I'm sure there is a general bias effect where the count of top seed winners is far greater than the count of top seed losers and therefore the average would tighten up anyway, but the data backs what we know anyway: Unbalanced teams get exposed.

Red Flag on Grant and the Irish this year, The Mayor and the Clones and less so, the Tarheels.
 
Joined
Aug 28, 2011
Messages
4,089
Reaction Score
5,890
Interesting stat insight in research today: I have to update my dataset, but from 2002 - 2011 top seeded teams (1-5) who lost in the first round had an average absolute difference of 40 in KenPom AdjO vs. AdjD rank. Top seeded teams who won in the first round have an avg difference of 20. Meaning the more unbalanced a team is on one side of the ball vs. the other, the more likely they are to get sniped.

Look at recent history (adjO, Adj D, Diff):

2014 Duke - (2, 116, 114) Ohio State - (128, 3, 125) VCU - (107, 6, 101)
2013 GTown - (78, 2, 76) Wisconsin ( 108, 1, 107)
2012 Duke - (10, 81, 71) Missouri (1, 146, 145)

Note these are post tourney ranks because I'm not a kenpom subscriber and I'm sure there is a general bias effect where the count of top seed winners is far greater than the count of top seed losers and therefore the average would tighten up anyway, but the data backs what we know anyway: Unbalanced teams get exposed.

Red Flag on Grant and the Irish this year, The Mayor and the Clones and less so, the Tarheels.

"from buried to bluebood"

^^^^^^^^^^^^^^^^

I like it VERY original.

Keep us updated with stats. My Kenpom expired on the 27th of last month. I can reup if you can crunch numbers to make a good thread! :) lol
 
Joined
Nov 21, 2012
Messages
4,634
Reaction Score
9,910
Creighton was also a huge disparity last year 3 Offense and 152 Defense and they got upset in Round 2 by a million.
 

OkaForPrez

Really Popular Poster
Joined
Aug 28, 2011
Messages
5,207
Reaction Score
26,712
OMG OMG!!!!!!!!! NO FREAKING WAY! RU SERIOUS??
lol I can't tell if your screwing with me but yea, I believed early it was going to be a special year. Dropped 100 cash on them too.
 

Inyatkin

Stairway to Seven
Joined
Jan 5, 2012
Messages
2,479
Reaction Score
9,715
Red Flag on Grant and the Irish this year, The Mayor and the Clones and less so, the Tarheels.
A Notre Dame team might underachieve? Impossible!
 
Status
Not open for further replies.

Online statistics

Members online
42
Guests online
1,299
Total visitors
1,341

Forum statistics

Threads
159,089
Messages
4,180,081
Members
10,048
Latest member
CT2SC


.
Top Bottom