Oversigning meets yppd

So, I can’t help my love for yppd (yards per per play differential).  I think it’s a great indicator of relative team strength.  I thought it would be interesting to compare a team’s yppd vs the size of their classes – so does oversigning help or hurt a team’s performance?

On the one hand, the data is clearly all over the place.  There are bad teams that sign a lot of players and there are good teams that sign a lot of players. 

That said, check this out – here are yppd scores (x-axis) mapped against the 4-year average size signing class (y-axis, using data from oversigning.com).  This is data for BCS schools only for 2006-2011 seasons.  The dots are specific “YPPD scores” (actual yppd + a SOS adjustment), the red line is a 5 data-point rolling average.

Interestingly, there is a very straight line here – the better the team is from a yppd perspective, the more players they sign on year on average.   Although, it trends up by only about 1.5 signees/yr (23 is at 0 yppd score [average] and about 24.5 as you move towards the top end).  

Also interestingly, is how, for the most part, top teams sign a high number of players.  Consider the list below – showing all teams with a 2 or greater yppd score – and how large their signing classes are.

So, the intuition that oversigning helps teams looks to be true, at least at a high level.



(the teams below are the top 27 yppd teams & avg # of signees over a 4 year period... the average class size is 24.7, the average class size of the bottom 27 is 22.8)
2006 West Virginia  29
2011 Alabama  28
2009 Alabama  28
2010 Auburn  28
2010 Arkansas  28
2010 Alabama  27
2007 Texas Tech  27
2011 LSU  26
2006 Arkansas  26
2008 Oregon  26
2011 Oregon  25
2010 Oregon  25
2007 Oklahoma  24
2007 West Virginia  24
2006 Texas Tech  24
2007 Mizzu  24
2008 Florida  24
2009 Florida  24
2007 Florida  24
2008 Oklahoma  24
2006 LSU  23
2007 LSU  23
2007 USC  23
2006 Florida  23
2011 Wisconsin  22
2009 Texas  21
2007 Ohio State  20

Does running = winning?

Here's how much a team rushes relative to what % of their games they win (2008-2011 data)

1st column = % of yards gained through running (for the whole season)
2nd column = % of games won

10% 32%
20% 42%
30% 49%
40% 51%
50% 55%
60% 56%
70% 54%
80% 42%

Note: the first column is actually a range, so 10%, really equals 10-19.9%

Then distribution of teams across the rushing categories by number of wins.

For teams that have won 10 or more games:

1st column = % of yards gained through running (for the whole season)
2nd column = % of teams in that bucket

10% 0%
20% 7%
30% 30%
40% 27%
50% 28%
60% 6%
70% 1%
80% 0%

Then teams with 7-9 wins

10% 1%
20% 7%
30% 31%
40% 37%
50% 16%
60% 4%
70% 3%
80% 1%

Then teams with 6 or fewer

10% 1%
20% 13%
30% 39%
40% 26%
50% 15%
60% 3%
70% 1%
80% 1%

A quick look at this tells me that you can be a very successful team with running or throwing the ball. Although, a higher proportion of very winning teams run the ball more.

ND runs the ball more with Kelly than it did with Weis, but it in the mid-to-high 30%'s - within the range in which successful teams commonly are.

Also - one more cut of data, teams winning 12 or more games over the last 4 years:

10% 0%
20% 8%
30% 32%
40% 20%
50% 36%
60% 4%
70% 0%
80% 0%

YPPD - Bowl edition

I’ve taken a few cuts of my yppd scoring applied to the upcoming bowl games – hopefully this is fun/interesting for some here.

Best bowls to watch

Maximizes both quality of the teams and closeness of the games
(#'s after the team are their 'yppd scores')

#10 - Orange Bowl Clemson 0.6 v West Virginia 1.3
0.9 yppd (#9 in quality), 0.8 gap (tied for #9)

#9 - Independence Bowl Missouri 0.7 v North Carolina 0.9
0.8 yppd (#11), 0.2 gap (tied for #3)

#8 - Sugar Bowl Michigan 1.4 v Virginia Tech 0.8
1.1 (#7), 0.6 gap (tied for #7)

#7 - Capital One Bowl Nebraska 0.9 v South Carolina 1.2
1.1 (#8), 0.3 gap (tied for #4)

#6 - TicketCity Bowl Houston 1.6 v Penn St. 1.1
1.4 (#5), 0.5 gap (tied for #6)

#5 - Outback Bowl Georgia 1.1 v Michigan St. 1.6
1.4 (#4), 0.5 gap (tied for #6)

# 4 - Champs Sports Bowl Florida St. 1.1 v Notre Dame 1.4
1.2 (#6), 0.5 gap(tied for #4)

#3 - Title game Alabama 2.9 v LSU 2.2
2.5 (#1), 0.7 gap (tied for #8)

#2 - Fiesta Bowl Oklahoma St. 1.7 v Stanford 1.4
1.5 (#3), 0.3 gap (tied for #4)

#1 - Rose Bowl Oregon 2.0 v Wisconsin 1.6
1.5 (#2), 0.3 gap (tied for #5)

Least Competitive Bowls
Games with the largest gaps in skill between the teams.

Cotton Bowl
Arkansas 1.3 v Kansas St. (0.4)

Texas Bowl
Northwestern (0.5) v Texas A&M 1.0

Gator Bowl
Florida 1.4 v Ohio St. 0.1

Las Vegas Bowl
Arizona St. (0.1) v Boise St. 1.0

New Mexico Bowl
Temple 0.1 v Wyoming (1.0)

Armed Forces Bowl
BYU 0.1 v Tulsa 1.1

Most competitive bowls
Games with the most closely matched teams

Holiday Bowl
California 0.6 v Texas 0.5

Pinstripe Bowl
Iowa St.(0.4) v Rutgers (0.4)

Kraft Fight Hunger
Illinois 0.5 v UCLA 0.5

Independence Bowl
Missouri 0.7 v North Carolina 0.9

GoDaddy.Com Bowl
Arkansas St. 0.5 v Northern Ill. 0.7

Famous Idaho Potato Bowl
Ohio 0.3 v Utah St. 0.6

Beef O’Brady’s
FIU (0.0) v Marshall (0.2)

New Orleans Bowl
La.-Lafayette (0.0) v San Diego St. (0.3)

Little Caesar’s Pizza
Purdue (0.7) v Western Mich. (0.4)

Other teams

Wisconsin




Washington


USC

LSU

Alabama

ND Team Performance since 2000

x-axis = yppd score
y-axis = win %
size of bubble = turnover margin (bigger = more favorable)