Ranking Bowl Game Watchability by FEI
SAE's Annual Mud Bowl: MVictors
The Fremeau Efficiency Index (FEI) isn't out yet for this week, so this data is outdated, but I had to turn in my office pool's bowl selections today for the ESPN confidence poll, and because I believe in using the internet to steal money from my co-workers, I grabbed their latest numbers (11/27) and did a little Excel work to make my picks.
In doing so, I realized my spreadsheet might also have a secondary, and perhaps more profitable use: determining a ranking of bowl games to watch so as to maximize my December/January bowl game enjoyment quotient without tripping the spousal "all you're doing is watching football; why don't you spend some time with me!" line (which with Misopogal is about 2.4 games per week).
All I did was create a list of bowl games, select the expected winner based on their FEI scores, and compare the standard deviations in FEI to decide which games were blowouts, tossups, etc. The spreadsheet's up on Google Docs if you want to play with it yourself.
The results are interesting enough I thought them worth sharing. Nebraska over Washington is a holy lock. Northwestern and Texas Tech might as well be playing themselves (although not really: see below).
In the following chart, the team listed first is expected to win. The FEI StD is the standard deviation between the combatants' FEI scores. Those at the top are the biggest expected blowouts; at the bottom are the pick-'ems. Michigan's expected to lose to Mississippi State, but it's only about a moderate expectation.
Bowl | Teams | FEI StD |
---|---|---|
Holiday | Nebraska vs. Washington | 0.18 |
Fiesta | Oklahoma vs. Connecticut | 0.14 |
Las Vegas | Boise State vs. Utah | 0.13 |
Military | Maryland vs. East Carolina | 0.12 |
Humanitarian | Northern Illinois vs. Fresno State | 0.11 |
Texas | Illinois vs. Baylor | 0.11 |
Little Caesars | Florida International vs. Toledo | 0.1 |
Music City | North Carolina vs. Tennessee | 0.09 |
Poinsetta | Navy vs. San Diego State | 0.08 |
Alamo | Oklahoma State vs. Arizona | 0.08 |
Capital One | Alabama vs. Michigan State | 0.08 |
St. Petersburg | Louisville vs. Southern Mississippi | 0.06 |
Cotton | LSU vs. Texas A&M | 0.06 |
New Mexico | BYU vs. UTEP | 0.05 |
GoDaddy.com | Middle Tennessee vs. Miami (OH) | 0.05 |
Outback | Florida vs. Penn State | 0.05 |
BBVA Compass | Pittsburgh vs. Kentucky | 0.05 |
Sun | Miami vs. Notre Dame | 0.05 |
BCS Championship | Auburn vs. Oregon | 0.05 |
Gator | Mississippi State vs. Michigan | 0.04 |
Kraft Fight Hunger | Nevada vs. Boston College | 0.04 |
Chick-fil-A | South Carolina vs. Florida State | 0.03 |
Liberty | Georgia vs. Central Florida | 0.03 |
Meineke Car | Clemson vs. South Florida | 0.02 |
Independence | Air Force vs. Georgia Tech | 0.02 |
Rose | Wisconsin vs. TCU | 0.02 |
Insight | Missouri vs. Iowa | 0.02 |
Sugar | Arkansas vs. Ohio State | 0.02 |
Armed Forces | Army vs. SMU | 0.01 |
Champs Sports | West Virginia vs. North Carolina State | 0.01 |
Orange | Stanford vs. Virginia Tech | 0 |
Hawaii | Hawaii vs. Tulsa | 0 |
New Orleans | Troy vs. Ohio | 0 |
Pinstripe | Syracuse vs. Kansas State | 0 |
TicketCity | Texas Tech vs. Northwestern | 0 |
Obviously this doesn't take into account things like having your entire offensive line return from injury, having your First Team All Big Ten Quarterback (Dan Persa: further proof that the dimension we are living in is not the real one) injured, teams that have historically put up embarrassing performances against a certain conference, or if your school's annual pre-bowl residence hall assault is liable to get a fifth of your team suspended*, so if you're planning on using this for your own pools, it's best you educate yourself on each teams' respective roster situations before making your selections.
I also added up the FEI scores of both opponents for each bowl game, to create an approximation of which bowls have the highest total performance, figuring games that feature better teams are more compelling. This quality of play index is pretty much in line with the general bowl ranks:
Bowl | Teams | Combined FEI |
---|---|---|
BCS Championship | Auburn vs. Oregon | 0.577 |
Orange | Stanford vs. Virginia Tech | 0.520 |
Sugar | Arkansas vs. Ohio State | 0.493 |
Rose | Wisconsin vs. TCU | 0.423 |
Capital One | Alabama vs. Michigan State | 0.418 |
Chick-fil-A | South Carolina vs. Florida State | 0.412 |
Champs Sports | West Virginia vs. North Carolina State | 0.387 |
Cotton | LSU vs. Texas A&M | 0.376 |
Insight | Missouri vs. Iowa | 0.361 |
Sun | Miami vs. Notre Dame | 0.283 |
Las Vegas | Boise State vs. Utah | 0.263 |
Fiesta | Oklahoma vs. Connecticut | 0.249 |
Meineke Car | Clemson vs. South Florida | 0.239 |
Alamo | Oklahoma State vs. Arizona | 0.230 |
BBVA Compass | Pittsburgh vs. Kentucky | 0.218 |
Liberty | Georgia vs. Central Florida | 0.172 |
Poinsetta | Navy vs. San Diego State | 0.155 |
Holiday | Nebraska vs. Washington | 0.153 |
Gator | Mississippi State vs. Michigan | 0.148 |
Kraft Fight Hunger | Nevada vs. Boston College | 0.129 |
St. Petersburg | Louisville vs. Southern Mississippi | 0.112 |
Outback | Florida vs. Penn State | 0.102 |
Pinstripe | Syracuse vs. Kansas State | 0.094 |
Music City | North Carolina vs. Tennessee | 0.091 |
Military | Maryland vs. East Carolina | 0.088 |
Independence | Air Force vs. Georgia Tech | 0.072 |
Texas | Illinois vs. Baylor | 0.069 |
Humanitarian | Northern Illinois vs. Fresno State | 0.000 |
Hawaii | Hawaii vs. Tulsa | -0.007 |
Armed Forces | Army vs. SMU | -0.047 |
TicketCity | Texas Tech vs. Northwestern | -0.049 |
Little Caesars | Florida International vs. Toledo | -0.127 |
New Orleans | Troy vs. Ohio | -0.154 |
New Mexico | BYU vs. UTEP | -0.179 |
GoDaddy.com | Middle Tennessee vs. Miami (OH) | -0.320 |
The data gave up some other interesting bowl tidbits:
- USC (28th, with an FEI of 0.113) is the best team that's not going to a bowl game.
- The best team not going due to things other than NCAA sanctions: Arizona State (41st, 0.061), those unlucky bastards.
- Miami (Not THAT Miami) is the worst team going to a bowl; their -0.198 is 110th out of 120.
If you subtract the (negative of the) defensive FEI of each team's opponent from its offensive FEI, then add up the scores, we get an approximation of which games will feature a lot of scoring. Top 10 predicted score-fests:
Bowl | Teams | OFEI v Opp |
---|---|---|
Military | Maryland vs. East Carolina | 0.76 |
Poinsetta | Navy vs. San Diego State | 0.59 |
Independence | Air Force vs. Georgia Tech | 0.56 |
Hawaii | Hawaii vs. Tulsa | 0.55 |
Alamo | Oklahoma State vs. Arizona | 0.48 |
BBVA Compass | Pittsburgh vs. Kentucky | 0.48 |
Gator | Mississippi State vs. Michigan | 0.47 |
BCS Championship | Auburn vs. Oregon | 0.37 |
Chick-fil-A | South Carolina vs. Florida State | 0.34 |
Armed Forces | Army vs. SMU | 0.28 |
Finally, for a Watchability Index, I ranked, then combined all three factors:
- Team Quality (TQ)
- High-Scoring (HS)
- Competitiveness (CO)
And came up with a final order of importance for games and how much they're worth watching, independent from school and conference interest:
Bowl | Teams | TQ | HS | CO | W.I. |
---|---|---|---|---|---|
Orange | Stanford vs. Virginia Tech | 2 | 19 | 3 | 84 |
BCS Championship | Auburn vs. Oregon | 1 | 8 | 18 | 81 |
Chick-fil-A | South Carolina vs. Florida State | 6 | 9 | 13 | 80 |
Sugar | Arkansas vs. Ohio State | 3 | 16 | 12 | 77 |
Rose | Wisconsin vs. TCU | 4 | 17 | 11 | 76 |
Hawaii | Hawaii vs. Tulsa | 29 | 4 | 1 | 74 |
Independence | Air Force vs. Georgia Tech | 26 | 3 | 8 | 71 |
BBVA Compass | Pittsburgh vs. Kentucky | 15 | 6 | 17 | 70 |
Gator | Mississippi State vs. Michigan | 19 | 7 | 15 | 67 |
Liberty | Georgia vs. Central Florida | 16 | 13 | 14 | 65 |
Alamo | Oklahoma State vs. Arizona | 14 | 5 | 25 | 64 |
Pinstripe | Syracuse vs. Kansas State | 23 | 18 | 4 | 63 |
Capital One | Alabama vs. Michigan State | 5 | 15 | 26 | 62 |
Poinsetta | Navy vs. San Diego State | 17 | 2 | 27 | 62 |
Armed Forces | Army vs. SMU | 30 | 10 | 6 | 62 |
TicketCity | Texas Tech vs. Northwestern | 31 | 11 | 5 | 61 |
Champs Sports | West Virginia vs. North Carolina State | 7 | 34 | 7 | 60 |
Insight | Missouri vs. Iowa | 9 | 30 | 9 | 60 |
Meineke Car | Clemson vs. South Florida | 13 | 35 | 10 | 50 |
Military | Maryland vs. East Carolina | 25 | 1 | 32 | 50 |
New Orleans | Troy vs. Ohio | 33 | 23 | 2 | 50 |
Kraft Fight Hunger | Nevada vs. Boston College | 20 | 24 | 16 | 48 |
Cotton | LSU vs. Texas A&M | 8 | 31 | 23 | 46 |
Sun | Miami vs. Notre Dame | 10 | 32 | 22 | 44 |
St. Petersburg | Louisville vs. Southern Mississippi | 21 | 20 | 24 | 43 |
Texas | Illinois vs. Baylor | 27 | 12 | 31 | 38 |
Las Vegas | Boise State vs. Utah | 11 | 27 | 33 | 37 |
Outback | Florida vs. Penn State | 22 | 28 | 21 | 37 |
Humanitarian | Northern Illinois vs. Fresno State | 28 | 14 | 30 | 36 |
Fiesta | Oklahoma vs. Connecticut | 12 | 29 | 34 | 33 |
New Mexico | BYU vs. UTEP | 34 | 21 | 20 | 33 |
Music City | North Carolina vs. Tennessee | 24 | 25 | 28 | 31 |
Holiday | Nebraska vs. Washington | 18 | 26 | 35 | 29 |
Little Caesars | Florida International vs. Toledo | 32 | 22 | 29 | 25 |
GoDaddy.com | Middle Tennessee vs. Miami (OH) | 35 | 33 | 19 | 21 |
Right, so of course all the good ones are on at the same damn time. Also, usefulness of this chart is made more questionable considering it doesn't take into account last week's games because of the old FEI scores, and things like Oregon/Auburn predicted as less competitive than Pitt/Kentucky. But it was fun, no?
----------------------------------------------------------
*Last year's 20-percent turnout beat the 2008 record of 18 percent; this year with sponsorship from Capital One, expect about 27 percent of Michigan State's football team to join in the campaign, with Greg Jones given the ultimate honor of being the marauder who gets to turn to the camera and say "What's in YOUR wallet?"
December 7th, 2010 at 7:58 PM ^
According to your predictions, the Big Ten is expected to finish 2-6 in the bowl games.
December 7th, 2010 at 8:10 PM ^
Sounds about right.
December 7th, 2010 at 8:16 PM ^
This is awesome, what a great read. Wish you luck in the office pool!
December 7th, 2010 at 9:21 PM ^
ain't stoppin' anyone, hoss.
December 7th, 2010 at 9:30 PM ^
One suggestion - when combining three factors as you've done above, rather than taking their ordinal rank and adding those together across factors, I recommend the following:
Take each value, then (1) subtract the smallest value and (2) divide by the range (i.e. difference between highest and lowest values).
Rather than an ordinal ranking system, you'll have something ranging from 0 to 100 that allows for greater variation. That is, if the difference between the highest-ranked and 10th highest-ranked in a category is significantly greater than the difference between the 50th and 60th ranked, that difference will be captured. This method provides more information than using a straight ordinal ranking.
December 7th, 2010 at 9:35 PM ^
Great idea. I thought about doing something like that right after I posted it. I'll probably do an update (and front-page it, which I kept this one off the FP because there was so much other content today) when the final FEI rankings are out.
December 7th, 2010 at 9:36 PM ^
This really points out how messed up all of the auto qualifiers are. At the least the bowls should pit top 50 teams against one another.
What would be better than seeing the top 8 or 16 play each other for the trophy. Maybe someday...
December 7th, 2010 at 9:59 PM ^
Dammit, man, that's the 4th time I've clicked on your avatar and thought "hmm...I'm gonna insert an image."
Maybe a pony?
December 8th, 2010 at 12:47 AM ^
Like the "This line intentionally left blank"
I do love the Desmond avatar...
Used to enjoy Magnus' but he has changed it...
December 7th, 2010 at 10:40 PM ^
Wow, great information. Will have to split the winnings from my office pool if it turns out profitable!
December 8th, 2010 at 12:23 AM ^
Two years ago, I used FEI* and kicked ass, something like 97th percentile in the ESPN.com fantasy game - and impressed my friends in the process.
Last year, I used FEI and finished well below the median...FEI got smoked last year. Of course I talked trash before bowl season about having a superior method, and ended up looking like a fool. While I still like FEI best out of the computer models, last year's bowl record did not help its credibility.
_____
*as Misopogon notes, some adjustments are necessary (injuries, obvious team motivational issues, etc.)
December 8th, 2010 at 9:05 AM ^
This analysis leaves out the "I don't care at all about this team and as such could never get behind this program enough to watch them compete on television" factor, which is huge for me in determining which games to watch.
Let's also not forget the "They have a football team?" factor as well (Tulsa, Troy, Ohio).
Else, altogether interesting break down.
Comments