FanPost

Anatomy of a Loss: What Stands Between a Team and the Playoffs?

Kevin Jairaj-USA TODAY Sports

Abstract:

In lieu of sitting and wringing my hands, I felt I'd take a stab at contributing the the college football playoff discussion beyond yelling at strangers on Twitter.

In trying to distance myself emotionally from the controversy at hand, I find myself time and time again coming back to the circular argument of loss comparison.

Granting that Florida State, Oregon, and Saban's Legion of Demons have all but secured for themselves three of four spots in the inaugural College Football Playoff, our comparison really comes down to three one-loss teams. The arguments for these teams, as you all know, range from "IT WAS A FLUKE WEEK" to "WE'VE IMPROVED" to Chubby kids yawning whilst holding up 61-58 towels in Waco. I noticed that despite the fact that the Playoff Committee serves the purpose of infusing common sense into the otherwise numerical and confusing BCS rankings, there seems to be a lack of rational decision making based on numerical comparison. I don't mean just "oh they won this game by this much" or "oh they didn't control the game" or any of that. I mean to suggest that we can somewhat responsibly determine (through common sense analysis of circumstance driven by quantitative logic) who should be in that fourth Playoff Spot, and so below is an attempt at using quantitative fixtures of qualitative data to aid us towards that answer.

If any of that sounds boring, run now.
_______________________________________________________________________________________________

From the CFP's website:

Selection committee members will have a wealth of information including review of video, statistics and their own expertise to guide them in their deliberations. They will emphasize obvious factors like win-loss records, strength of schedule, conference championships won, head-to-head results and results against common opponents. The playoff group has retained SportSource Analytics to provide the data platform for the committee’s use. The platform will allow the committee members to compare and contrast every team on every level possible.

It should be noted that the committee will not use a single data point such as the Ratings Percentage Index (RPI) that is used for NCAA championships.

From this we can pull the following list of factors going into the rankings:

1. Win Loss

2. SOS

3. Conference Championships

4. Head To Head Results

5. Common opponents

There are a lot of ways you could go with this argument, but I am going to focus on #1.

In comparing teams' won-loss record, everyone internally and subconsciously creates an equation of significance. The human mind is selective, spotty, and extremely biased, so each person collects whatever data best suits their purpose, ignores the rest, and then spins the argument towards themselves. It's easy to just count only head to head (Baylor) and conveniently ignore facts like two score losses to unranked teams.

Instead of allowing that subjective and subconscious derivation of value drive the discussion, why not reason through the category with quantitative coefficients?

So, what this discussion really comes down to, in my mind, is the loss that each team has, if only for the reason that without that loss, any of these three teams is an undefeated conference champion and obviously in the playoff, no questions asked. The decision really comes down to the following three games:

NR Virginia Tech 35, #8 Ohio State 21

NR West Virginia 41, #4 Baylor 27

#5 Baylor 61, #9 TCU 58

These three games are all that separate each team from an unquestionably merited playoff birth. So, as a result, we can compare these losses based on circumstance. Each loss must be considered in light of a few variables, each with a different weight. So, each loss is some function of the following: Result, Margin of Victory, Location, Injuries, External Influence, and Quality of Opponent. The data is not big enough to run a significant least squares regression, so our conversation will me a little... methodologically interesting.

DISCLAIMER: THIS IS A LITTLE FLUKY - I'm more attempting to use the numbers as a point for then having a discussion. Let's move forward.

Let's say, for the sake of continuity, that the perfect loss, the loss least damaging to a college football team's record, is a loss to a better team by a close margin on the road with some external factors (i.e. officiating) playing a major factor. (You can see where I'm headed with this.)

Granted, all of the below is extremely arbitrary.

Result: 40 - a win in this game is worth 40, a loss worth zero.

Margin of Victory: subtracted from total score.

Location: 10 (dummy variable - home is 0, 10 is on the road, 5 is neutral site).

Injuries: 10 points for key injuries (i.e. experience starters), 0 for a completely healthy squad.

External Influence: 10 points for a game altering call, 0 for no major hi-jinks

Quality of Opponent: A sliding scale of 0-15 based on the opponents ranking - 0 points out of top 25, with 5 points being awarded for the last ten teams of the top 25, 10 points awarded for a team 15-11, and all 15 points being awarded for an opponent in the top ten.

I decided post hoc to eliminate "external influence" because it provided an opportunity for me to be nothing but subjective for my own purposes, and that's dishonest.

Now, let's create a table where we assign values to each of the categories for each of the games, and then use that as a reference point for discussion.

Variables

NR Virginia Tech 35, #8 Ohio State 21

NR West Virginia 41, #4 Baylor 27

#5 Baylor 61, #9 TCU 58

Result W=40, L=0

0

0

0

M.o.V 1-15

-14

-14

-3

Location H=0, Road= 10, Neutral=5

0 (weather was good)

10 (it was freezing)

8 (weather was good)

Injuries 1-10

10 – W/O Starting QB in 2nd game of season

0 – Mostly Healthy

0 – Mostly Healthy

Quality of Opponent

NR = 0, #25-#16 = 5, #15-#10=10, #9 on = 15

0 – Unranked Va Tech (6-6)

0 – Unranked West Virginia (7-5)

15 - #6 Baylor (11-1)

TOTAL GAME SCORE

-4

-4

20

Now, granted, this is a work in progress, and perhaps extremely arbitrary, as noted before, but, let's just use this as a definition of circumstance. (In the comments, I'd love for some methodological critique, because we could do something very cool here.)

These three teams are all but equal in the media's eyes. Three 11-1 teams, separated from the playoff only by the above losses. In my mind, the degree of separation from the playoffs should be determined by the only variable that is keeping them out of the playoffs, their above loss. Playoff teams don't lose to unranked teams.

Theoretically, this could be extrapolated for the whole season, and if I get really incensed later today, perhaps I will do that. Each team's playoff merit comes down to a function of the circumstances of their wins and losses, and with such close comparison of one loss teams, the quality of the loss has to be the defining factor. It is literally the barrier between three good teams and the playoffs, and the gravity of that barrier should be the defining factor in deciding who gets in.

edit= for fun, I added ALABAMA and OREGON's only losses, but this matters less because they were both conference champions and media darlings. Also, I added a poll, for click bait.

Variables

NR Virginia Tech 35, #8 Ohio State 21

NR West Virginia 41, #4 Baylor 27

#5 Baylor 61, #9 TCU 58

NR Arizona 31, #2 Oregon 24

#11 Ole Miss 23, #3 Alabama 17

Result W=40, L=0

0

0

0

0

0

M.o.V

-14

-14

-3

-7

-6

Location H=0, Road= 10, Neutral=5

0

10

10

0

10

Injuries 1-10

10 – W/O Starting QB in 2nd game of season

0 – Mostly Healthy

0 – Mostly Healthy

0 - Mostly Healthy

0 – Mostly Healthy

Quality of Opponent

NR = 0, #25-#16 = 5, #15-#10=10, #9 on = 15

0 – Unranked Va Tech (6-6)

0 – Unranked West Virginia (7-5)

15 - #6 Baylor (11-1)

15 - #7 Arizona (10-2)

10 - #12 Ole Miss (9-3)

TOTAL GAME SCORE

-4

-4

22

8

14

All comments, FanPosts, and FanShots are the views of the reader-authors who create them.