Untitled

Tennessee High School Football Ratings with and without the Margin of Victory

By Earl Nall (Bio)

The collegiate Bowl Championship Series (BCS) formula is being modified to eliminate the margin of victory (MOV) component in the computer ratings. I modified my Tennessee high school football computer ratings to see what affect eliminating the MOV would have on them. From a statisticians view this was a good experiment as my ratings terminate at the end of the regular season. Since Tennessee has a playoff system I can always determine how good the ratings were by looking at the results of the play-offs and compare them to my final ratings. This is something the NCAA cannot do since they do not have a play-off system and can’t really determine how good the BCS is at picking the top teams.

I have been doing ratings for ten years and they consistently hit on 80-84% of the games during the last weeks of the season. Also, never has a team won the state championship that was not ranked in the top 10 at the end of the season.

Like a lot of computer rating systems I use a cap on the MOV. I cannot imagine not using the MOV. The margin of victory is the single most important statistic in determining a team’s strength. If a team wins a game 45-0, they are almost always going to win the statistical game too. In all my studies in determining what are the important factors in a team’s strength, MOV comes out on top. If in the final score one could eliminate points off turnovers, substitutes, and trick plays - computer ratings could be unbelievably accurate.

To see how not using MOV could cause problems – look at this scenario. Suppose you have a ten-team conference and all the teams play each other. Going into the last week of the season Team A has defeated all its opponents handily by a score of 55-0 – inserting second and third teams to finish each game. For the last game of the season they are going to play Team B who is also undefeated, but has won each of its games 7-6. The rest of the schools in the conference have beaten up on each other. So for the common opponents Team A has won by a margin of 54 points more than Team B has.

Going into the final game, common sense would tell you that Team A is going to defeat Team B. In fact, all the voting polls would have Team A solidly ranked No. 1 and Team B a distant No. 2. But the computer systems would have to have these teams’ dead even since you can’t count MOV and since all the teams in the conference have beaten up of each other so there is no clear team to put into 3^rd place.

So, how did the Tennessee high school ratings do using MOV and without MOV?

In Tennessee there are eight state championships, so 16 teams make it to the final game. Brackets are pre-set before the season starts. Using my current system which uses the MOV, 13 of the 16 teams that were in the championship game were “picked to be there” by the ratings. This is a darn good number I think. You can say you don’t like computer ratings, but historically good computer ratings program will have that kind of accuracy. When the MOV was not used, the ratings picked 9 of the 16 teams to make the finals. While not bad, it is still just a little over 50% while using the MOV was over 80%.

Here is the breakdown by class:

Final Regular Season Ratings
Class 1A
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	Ezell-Harding (Champions)	1	Ezell-Harding (Champions)
2	Univ. School Jackson	2	Collinwood
3	Christ Pres. Academy	3	Bruceton Central
4	Collinwood	4	Gordonsville
5	Bruceton Central	5	Univ. School Jackson
6	Jo Byrns	6	Christ Pres. Academy
7	Cloudland (Runnerup)	7	Cloudland (Runnerup)
8	DCA	8	Midway
9	Gordonsville	9	Greenback
10	Boyd-Buchanan	10	Cascade

Teams in bold are the highest rated teams that could reach the finals based on the brackets
Example: Teams 1-6 were in the west bracket this year, while Cloudland was the highest ranked team from the east

Notes: Ezell-Harding actually finished second in the final ratings using MOV, but just by a point.
Since they played University School of Jackson at home in the play-offs they were
actually favored since they got the 2.5 point home field advantage.

Final Regular Season Ratings
Class 2A
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	Westview (Runnerup)	1	McKenzie
2	David Lipscomb	2	Westview (Runnerup)
3	Goodpasture (Champions)	3	Alcoa
4	Tyner	4	Lewis County
5	Lewis County	5	Tyner
6	McKenzie	6	David Lipscomb
7	Union City	7	South Greene
8	Huntingdon	8	Goodpasture (Champions)
9	Humboldt	9	Sweetwater
10	Milan	10	Booker T. Washington

Teams in bold are the highest rated teams that could reach the finals based on the brackets

Final Regular Season Ratings
Class 3A
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	Portland	1	Portland
2	Covington	2	Anderson County
3	McNairy Central	3	McNairy Central
4	Trousdale County	4	Covington
5	Anderson County	5	Trousdale County
6	Crockett County (Runnerup)	6	Sycamore
7	Austin-East (Champions)	7	Austin-East (Champions)
8	Marshall County	8	Knoxville Fulton
9	Knoxville Fulton	9	Kingsbury
10	Smith County	10	Smith County
11		11	Crockett County (Runnerup)

Teams in bold are the highest rated teams that could reach the finals based on the brackets


Final Regular Season Ratings
Class 4A
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	Maryville (Champions)	1	Tennessee High
2	Hillsboro (Runnerup)	2	Dyersburg
3	Cleveland	3	Hillsboro (Runnerup)
4	Red Bank	4	Morristown West
5	Memphis East	5	Columbia
6	Morristown West	6	Maryville (Champions)
7	East Ridge	7	East Ridge
8	White House	8	White House
9	Columbia	9	Memphis East
10	Tennessee High	10	Pearl-Cohn

Teams in bold are the highest rated teams that could reach the finals based on the brackets


Final Regular Season Ratings
Class 5A
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	Riverdale (Champions)	1	Riverdale (Champions)
2	Kingsport Dobyns Bennett	2	Kingsport Dobyns Bennett
3	Hendersonville (Runnerup)	3	Hendersonville (Runnerup)
4	Brentwood	4	Brentwood
5	Germantown	5	Whitehaven
6	Collierville	6	Germantown
7	Ooltewah	7	Houston
8	Gallatin	8	Collierville
9	Whitehaven	9	Hunters Lane
10	Hunters Lane	10	Central-Merry


Teams in bold are the highest rated teams that could reach the finals based on the brackets

Final Regular Season Ratings
Division II AAA
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	Bishop Byrne (Champions)	1	Bishop Byrne (Champions)
2	SBEC (Runnerup)	2	SBEC (Runnerup)
3	Rosemark	3	St. Andrews
4	FACS	4	FACS
5	St. Andrews	5	Rosemark
6	Memphis Catholic	6	Elliston
7	Elliston	7	Memphis Catholic
8	Kings Academy	8	Kings Academy

Teams in bold are the highest rated teams that could reach the finals based on the brackets


Final Regular Season Ratings
Division II AA
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	Briarcrest (Runnerup)	1	Harding Academy
2	BGA (Champions)	2	Briarcrest (Runnerup)
3	Knoxville Webb	3	BGA (Champions)
4	Harding Academy	4	Knoxville Webb
5	ECS	5	ECS
6	Notre Dame	6	Notre Dame
7	Knoxville Catholic	7	Knoxville Catholic
8	St. Benedict	8	St. Benedict

Teams in bold are the highest rated teams that could reach the finals based on the brackets


Final Regular Season Ratings
Division II A
Using Margin of Victory		Eliminating Margin of Victory
Rank	Team	Rank	Team
1	McCallie (Champions)	1	McCallie (Champions)
2	Brentwood Academy (Runnerup)	2	Brentwood Academy (Runnerup)
3	MUS	3	MUS
4	MBA	4	CBHS
5	Baylor	5	MBA
6	CBHS	6	Baylor
7	Father Ryan	7	Father Ryan

Earl Nall has been doing computer ratings for Tennessee high school football for ten years. He majored in mathematics and physics in college and his mathematics professor was Herman Matthews whose ratings for Scripps-Howard are part of the Bowl Championship Series (BCS). Mr. Nall worked for years as a statistician and computer programmer in Oak Ridge, Tennessee. He is now the technology coordinator for the Tennessee Secondary School Athletic Association (TSSAA), which oversees high school sports in the state. He lives in Kingston, Tennessee with his wife and daughter who is a college sophomore.