Long ago, in the early 90's, I was a sportswriter. And a gambler. My sports writing aided my efforts in gambling because I had access to all kinds of statistics that the average person didn't. I could get on the agate wire and print out up-to-the-minute rankings that mattered - chief among them, rush offense, rush defense, third down conversion rates, and schedule strength. Back then, I didn't need much more than those numbers to do very well in the college game. Because after about game four, some things don't change. A team that cannot defend the run will continue not to do so, because there are no free agents to sign. Winning a bet usually required little else than hawk-eyeing the bottom of the rush defense category and waiting for them to line up against a good rushing team with a spread under 10.
Now, it isn't as easy, because everyone has access to these stats. They are easily found at ESPN, CBSsports, hell, at the NCAA's very own site. So, astute gamblers have had to evolve, to look beyond the simplicity of basic numbers to statistics that are far more esoteric. For years now, I have placed a great emphasis on just a few stats: yards per play, yards per play allowed, yards per point, yards per point allowed, yards per pass attempt, and schedule strength. Let's take a look at Saturday night's matchup using these numbers.
Here's the raw data.
Team YPP YPPAlw YPPT YPPTAlw YPA SS
LSU 5.1 3.2 9.6 17.3 7.7 42
WVU 6.1 4.9 10.3 16.9 8.6 119
(Forgive the formatting issue....)
The yards-per-play number show an advantage for WVU of a yard, but that is mitigated by the fact that LSU's defense is allowing 1.5 yards less per play than the Mountaineers. The lower the yards-per-point number is, the better, as it is a metric that reveals the efficiency of a team's offense. Conversely, the higher a team's yards-per-point allowed, the better that team's defense. As you can see, LSU is just slightly better than WVU in both of those categories. I have always been a proponent of class ratings in horse racing, and Sagarin schedule strength provides that component in college football. Again, LSU has an advantage, and a pretty decent one as the Tigers are nearly in the top third of the FBS, while the Mountaineers currently reside in the bottom third.
The main way I use these numbers is by looking at the differentials that they present.
LSU 's offense has an advantage of .2 yards per play, while WVU's offense may struggle with a defense that is yielding just 3.2 yards per play. The yards-per-point differentials between each team's offense and defense are pretty much a wash, but the edge there has to go to LSU due to its substantial class advantage in teams played.
The obvious thing is to do what I did above - lining up one team's offensive numbers against its opponents defensive ones, but for a few years now, I have put more faith in comparing a team's offensive numbers to its own defensive ones, a la Bud Goode's Killer Stat for pro football (which looks only at yards per attempt, since most NFL teams run the ball about the same...I use yards per play in college football because running games are not created equal at this level). LSU has a net yards-per-play of 1.9, to WVU's 1.2, and it wins on the yards-per-point comparison as well with a -7.7 to the Mountaineers -6.6 (the lower the net number, the better as it indicates a low yards-per-point and a high yards-per-point allowed).
The stone-cold objective gambler in me thinks that it might be too difficult of a chore for the Mountaineers to win this game, as the only real advantage they have is about a yard better number in yards per attempt. I certainly would not bet the farm. And yet, the games aren't played on paper, and all this data notwithstanding, I think the Mountaineers win on the strength of the fact that Holgorsen's offenses have never been held in check by top-ranked defenses.
I am sticking with my earlier prediction: WVU 27, LSU 21.