A Statistical Case for Why Manti Te'o Should Have Won the Heisman Trophy

The 2012 Heisman Trophy has been awarded to Texas A&M quarterback Johnny Manziel, and it wasn't exactly an unjust decision. Manziel broke an SEC record with over 4,600 total yards of offense after all, and he led Texas A&M to a surprising 10-2 season, complete with a dramatic win at No. 2 Alabama.

Dude had a hell of a year.

The runner-up to the award was Notre Dame linebacker Manti Te'o, who received the second-highest amount of votes for a Heisman runner-up, missing out on the top spot by about 300 points. That's a remarkable outcome, especially since a full-time linebacker had never come that close to winning the Heisman before (John David Crow played halfback and linebacker in his Heisman-winning campaign of 1957).

So while we celebrate the wonderful performance of the linebacker we endorsed for the Heisman Trophy, we can't help but figure that Te'o's case was actually undersold by Notre Dame.

It's difficult to quantify quality performance for an individual defender in football. Oh, there's plenty of data available, mind you—this isn't the same quandary of the stat-less, thankless offensive lineman—but just because someone gets a tackle doesn't mean a positive defensive result has just been achieved. Completed passes usually result in tackles, while incomplete passes never do, for example. Ask any coach which he'd rather have.

Moreover, as we explained last time, a high number of tackles doesn't necessarily mean the defense is doing a great job; in fact, it probably means the opposite. Houston's defense had four defenders with more tackles than Te'o in 2012. Houston's defense is ranked 115th out of 120 teams by the NCAA.

Do you honestly think Te'o would be the fifth-best player on that defense?  

Things like tackles-for-loss are easier quantifiers, but they're role-specific. An outside linebacker is generally more likely to have tackles-for-loss than a middle linebacker, for example; but that hardly makes outside linebackers better defenders than middle linebackers by necessity. Similarly, interceptions are great results for cornerbacks, but the best cornerbacks don't get passes thrown their way—and the bad ones get targeted much more often.

So again: For individual defenders, it's hard to quantify quality.

At the very least, what we can do is not look just at tackles as a raw stat, but individual tackles in proportion to other stats. If a defender has a relatively high percentage of his team's tackles—or a relatively high percentage of tackles per defensive plays as a whole—then the importance of his role in the team's defense can be more clearly defined. It's not perfect, but nothing in college football is. Well, nothing except the CBS theme.

Better yet, combine that with looking at the yardage given up by a defense and plays per game allowed, and we can see whether that defender is playing a major role on a great defense or a bad one. A linebacker who makes 15 percent of his team's tackles on a defense that can't crack the top 100 nationwide isn't automatically a bad linebacker, but he's clearly not dominant either.

So, we took a look at the leading tacklers on every single defense in the nation, thanks to the stats at NCAA.org. And even on a loaded defense like Notre Dame's, Te'o still stands out on a much higher level than just plain total tackles would indicate.

One note: Te'o does not lead the nation in tackles per defensive play, nor in percentage of his team's overall tackles. On these metrics, however, Te'o is at least ranked much higher than 59th in the nation.

In terms of a percentage of his team's overall tackles, Te'o makes the leap to 31st in the nation. Among top-10 defenses (for the sake of fitting onto a page without overwhelming it), he's still fourth:

Rank	Team	Games	Tackles	Team TPG	Leading Tackler	Tackles	Tackle %
1	Alabama	13	832	64	C.J. Mosley	99	11.90%
2	Florida State	13	815	62.69	Christian Jones	85	10.43%
3	BYU	12	746	62.17	Brandon Ogletree	92	12.33%
4	Michigan State	12	733	61.08	Max Bullough	102	13.92%
5	Florida	12	702	58.5	Josh Evans	79	11.25%
6	Notre Dame	12	796	66.33	Manti Te'o	103	12.94%
7	Bowling Green	12	761	63.42	Gabe Martin	69	9.07%
8	LSU	12	819	68.25	Kevin Minter	111	13.55%
9	Boise State	12	873	72.75	J.C. Percy	101	11.57%
10	UConn	12	849	70.75	Yawin Smallwood	120	14.13%

Percentage of overall tackles is problematic, though, since the way they're compiled is problematic.

Scorekeepers can vary from one school to another in terms of what they consider an assisted tackle, and each school has its own scorekeeper. Thus, we get a situation where Kansas' 113th-ranked defense had 544 solo tackles and 295 assisted tackles...and our old friend Houston had 469 solo tackles and 668 assisted tackles. That goes far beyond mere stylistic differences in defense.

What's less subjective than counting assisted tackles is counting plays, and when we look at individual tackles as a percentage of total plays, Te'o rises even further among team leaders, to 24th in the nation.

Now, among top-10 defenses, only Minter and Smallwood rank better:

Rank	Team	Games	Plays	Def. Plays Per Game	Leading Tackler	Tackles	Tackles Per Play (in %)
1	Alabama	13	782	60.15	C.J. Mosley	99	12.66%
2	Florida State	13	849	65.31	Christian Jones	85	10.01%
3	BYU	12	722	60.17	Brandon Ogletree	92	12.74%
4	Michigan State	12	757	63.08	Max Bullough	102	13.47%
5	Florida	12	789	65.75	Josh Evans	79	10.01%
6	Notre Dame	12	757	63.08	Manti Te'o	103	13.61%
7	Bowling Green	12	758	63.17	Gabe Martin	69	9.10%
8	LSU	12	797	66.42	Kevin Minter	111	13.93%
9	Boise State	12	811	67.58	J.C. Percy	101	12.45%
10	UConn	12	811	67.58	Yawin Smallwood	120	14.80%

The entire list of leading tacklers and other relevant statistics is here if you'd like to peruse it. It's locked for editing, so no, you may not change your favorite team's leading tackler to Bobby Boucher and give him 637 tackles on the year (although that would be funny). 

Moreover, tackling is just one aspect of playing defense. There's also taking the ball away, and there Te'o excelled as well. As has been mentioned numerous times, Te'o shared the FBS lead in takeaways with Iowa State defensive back Durrell Givens with nine on the year. What's more, Te'o got seven of his nine takeaways from interceptions, which are less subject to chance than fumble recoveries.

Also, this isn't a statistical argument, but one of those picks was this game-clinching pick at Oklahoma, which remains one of Notre Dame's biggest wins of the season:

And last, it's worth reiterating that statistically, Notre Dame had a great defense, and Te'o was the centerpiece of that great defense. Would he have been a Heisman contender if he were playing for, say, Colorado State? Of course not, but neither would have anybody else. The last Heisman contender to play for a wretched team was Troy Davis at Iowa State, and that was back when cracking 2,000 yards rushing really meant something.

So Te'o is a top-25 tackler in terms of his team's overall production (and that is a team full of defenders willing to wreck a play) and a top-two takeaway producer—all on an elite defense of an undefeated team playing for a national championship. 

That's a strong resume.

It's stronger than the "best player on the best team" meme that usually gets tossed around this time of year. And if college football had considered tackling stats as percentages instead of raw numbers (to say nothing of the fact that Te'o had two missed tackles all year, according to ESPN during its Heisman Trophy broadcast—where was that stat in the conversation before then?), perhaps Te'o would have been more widely accepted as a potential Heisman Trophy winner.

And in a race that came down to about 300 voting points, that might have swung the difference on who ended up raising the Heisman Trophy on Sunday. Alas.