FPA: another reason why Georgia is doomed in 2010

I’ve really been digging into the statistical analysis that Football Outsiders‘ Bill Connelly and Brian Fremeau have done with their 2010 College Football Almanac (a bargain that you can download for a mere five bucks, by the way).

One metric they’ve devised that’s gotten my attention – and you’ll see why in a minute – is something they call Field Position Advantage (FPA).  Basically, they’ve determined the scoring value of every yard line on the field (how much an average offense can be expected to score against an average defense from that place).  They then take a measure of a team’s entire set of offensive series for each game over the season and factor in non-offensive scoring.  For a particular game, the sum of each team’s scoring value is combined and the ratio of one team’s scoring value to the total is FPA.

As they put it,

… FPA is a description of which team controlled field position in the game and by how much.  Two teams that face equal team position over the course of a game will each have an FPA of .500.  Winning the field position battle is quite valuable.  College football teams that play with an FPA over .500 win two-thirds of the time.  Teams that play with an FPA over .600 win ninety percent of the time.

You can see where I’m heading with this, can’t you, Dawg fans?  Here’s a stat that quantifies a lot of stuff that drove us nuts last year:  turnover margin, penalties, directional kicking, Logan Gray fair catches… you name it.  It’s kind of like a Unified Theory of Mediocrity.  (It’s also going to account for positives like Boykin’s returns and Butler’s net punting work.)

And when you look at the FPA rankings for the SEC last season, it confirms your worst suspicions.  [Note:  Numbers in parenthesis are national standings.]

Florida, 0.547 (11)

LSU, 0.537 (15)

Alabama, 0.527 (23)

Arkansas, 0.523 (29)

Kentucky, 0.517 (35)

Tennessee, 0.509 (45)

Vanderbilt, 0.509 (45)

Mississippi State, 0.504 (56)

Auburn, 0.499 (62)

Mississippi, 0.494 (65)

South Carolina, 0.476 (89)

Georgia, 0.466 (104)

Not pretty.  But not in the least bit surprising, either.

I pestered Bill for some specific context on Georgia’s FPA.  He was kind enough to respond by e-mail with this:

Here’s a little chart showing how teams in different FPA ranges tend to fare in terms of win percentage:

FPA Range	BCS  Teams	Non-BCS	ALL
.400 to .429	0.226	0.205	0.212
.430 to .459	0.303	0.285	0.293
.460 to .489	0.473	0.378	0.421
.490 to .519	0.559	0.484	0.526
.520 to .549	0.695	0.615	0.672
.550 to .589	0.730	0.750	0.738

Extremely linear, which is always what you want to see when talking about the legitimacy of a measure.  Basically BCS teams falling into Georgia’s range (.460 to .489) usually have about a .473 win percentage, which is about 5.6 wins over a 12-game season. [Emphasis added.] Of the ten BCS teams who have had an FPA between .460 and .470 in the last three years, only two had winning records (Georgia 2009 and NC State 2007) and six went 4-8 or worse (Colorado 2009, Purdue 2008, Iowa State 2007, Indiana 2008, Syracuse 2009, Texas A&M 2008).  It is VERY rare for a good team to be that bad in terms of Field Position Advantage.

You can look at this data in one of two ways:  either the Dawgs overachieved last season, or they were a talented bunch that dragged themselves down because they didn’t do a lot of the little things well (or did some of them stupidly, if you prefer).  I lean towards the latter, not because I’m being a homer, but because I have a hard time putting a Martinez-coached defense and overachievement in the same thought pattern.

Now, Bill is careful to warn that you can’t read too much into FPA – specifically, that it’s not a very predictive tool going from one season to the next.  Given that turnover margin is generated fairly randomly year to year, that makes a good deal of sense.

Even so, if you think about a few things going for Georgia in 2010, such as swapping Jon Fabris for Warren Belin on special teams, an easier schedule and our old buddy regression to the mean, it’s hard not to see an improvement coming in FPA.  If the offense and defense do no better than hold their ground from last season, that would still indicate a decent possibility that better days lie ahead in the won/loss department.  At least that’s my story and I’m sticking to it for now.