clock menu more-arrow no yes mobile

Filed under:

Know how baseball and basketball have deep stats for individual players? Here’s a cool way to do that in football.

This is how numbers can tell us more about what a player does well.

If you buy something from an SB Nation link, Vox Media may earn a commission. See our ethics statement.

UCLA v USC Photo by Harry How/Getty Images

Football doesn’t lend itself toward the clear individual stats we see in sports like baseball. A quarterback has passing yards and a completion rate, but he needs an offensive line to block for him and a player to catch his pass. A defensive end has tackles and TFLs, but play-by-play data doesn’t provide us with much context — how many opportunities did he really have? And how many times did he risk a big gain in the name of a TFL?

We’re never going to have a VORP for college football players, in other words, at least not without using the type of player grades that make coaches uncomfortable.

All individual stats in football are like RBI in baseball or plus/minus in basketball: useful but lacking context.

That said, we can still keep adding descriptiveness to the player data we’ve got.

At some point in this offseason, I will finish a redesign of my S&P+ ratings. Its team explosiveness and efficiency measures will begin factoring in expectations, based on down-and-distance and field position. (For example, gaining 18 yards on third-and-19 shouldn’t be judged less efficient than gaining three on third-and-3; the inefficient stuff happened on first or second down, not third.)

This concept can also be used in a pretty exciting way for individual stats. Here are my new individual expected efficiency and explosiveness measures, what I call Marginal Efficiency and Marginal Explosiveness.

  • Marginal Efficiency: the difference between a player’s success rate* (passing, rushing, or receiving) or success rate allowed (for an individual defender) and the expected success rate of each play based on down, distance, and yard line.
  • Marginal Explosiveness: the difference between a player’s IsoPPP** (passing, rushing, or receiving) or IsoPPP allowed (for an individual defender) and the expected IsoPPP value of each play based on down, distance, and yard line.

For offensive players, the larger the positive value, the better. For defensive players, it’s the opposite — the more negative, the better.

* Success rate: a common Football Outsiders tool used to measure efficiency by determining whether every play of a given game was successful or not. The terms of success in college football: 50 percent of necessary yardage on first down, 70 percent on second down, and 100 percent on third and fourth down.

** IsoPPP: the average equivalent point value of successful plays only.

Let’s walk through some examples.


UCLA quarterbacks: Marginal Efficiency

Category Josh Rosen Devon Modster
Category Josh Rosen Devon Modster
Completion Rate 62.6% 64.6%
Yards per Completion 13.3 13.2
Marginal Efficiency 9.8% 2.7%
Marginal Explosiveness 0.19 0.31

When we look at a passer’s stat line, we tend to think completion rate equals efficiency. It’s just close enough to pull it off. A guy completed 70 percent of his passes in a game? He was efficient! Fifty-five percent? Not so much!

But that doesn’t really tell us if those completions were of any use. A quarterback has the option of completing a three-yard pass just about any time he wants, but that won’t really take his offense anywhere.

In 2017, UCLA’s Josh Rosen completed a solid 63 percent of his passes. He missed a couple of games, however, and backup Devon Modster, a redshirt freshman, completed 65 percent. Modster’s sack rate (3.7 percent) was lower than Rosen’s, too (5.4 percent). Good, right?

Not really. Modster completed 67 percent of his passes against Utah, but the Bruins went three-and-out seven times and lost, 48-17. He completed 62 percent of his passes against Kansas State in the Cactus Bowl, but despite scoring two touchdowns in three plays in the second quarter, the Bruins scored 17 in an 18-point loss.

Marginal Efficiency tells a clearer tale than completion rate. Based on down and distance, Rosen’s success rate was nearly 10 percent higher than expectation; Modster’s was 7.1 percent lower.

That might not sound like a lot until you realize 7.1 percentage points separated the No. 11 FBS offense in success rate in 2017 (USC, at 46.6 percent) and the No. 98 offense (Southern Miss, at 39.5).


NCAA Football: Army at Navy
Darnell Woolfolk
Bill Streicher-USA TODAY Sports

Army rushing: Marginal Efficiency

Category A. Bradshaw (QB) D. Woolfolk (FB) A. Davidson (FB) K. Walker (SB)
Category A. Bradshaw (QB) D. Woolfolk (FB) A. Davidson (FB) K. Walker (SB)
Rushes 242 157 116 86
Opp. Rate* 49.2% 40.1% 34.5% 45.3%
Hlt Yds/Opp** 7 3.4 4.9 8
Marginal Eff. 10.9% 6.1% 1.1% 1.6%
Marginal Exp. 0.01 -0.21 -0.21 0.14

* Opportunity rate: the percentage of carries in which the offensive line “does its job” and produces at least five yards of rushing.

** Highlight yards per opportunity: on carries that gain at least five yards, the portion of the yardage credited to the running back, per the Football Outsiders line yardage formula.

For a while now, I have used Opportunity Rate and Highlight Yards per Opportunity as ways of defining efficiency and explosiveness for rushers, and I maintain those are useful tools. But a single, huge run can skew your highlight yardage, and your explosiveness potential is dictated by where you take your hand-offs.

To look at how Marginal Efficiency can define a runner, let’s look at one of the runningest backfields in the country: Army’s.

In a triple option, the roles are pretty defined. The fullback’s job is to hammer away between the tackles, succeeding just enough to suck the defense in and open up lanes for the speedy slot backs on the edge. And depending on the quarterback’s style, he might end up with some grind-it-out stats or those befitting a slot back.

For the super-efficient Army attack — the Black Knights were second in FBS in overall success rate — the QB was pretty explosive, in part because of success on the interior. Fullbacks Darnell Woolfolk and Andy Davidson, 235-pounders each, ground out five yards per carry between the tackles, with Woolfolk the more efficient.

Primary slot back Kell Walker averaged an excellent 7.3 yards per carry on the perimeter, with 45 percent of his carries gaining at least five yards. Predictably, that opportunity rate was quite a bit higher than that of the fullbacks.

Marginal Efficiency, however, adds a layer of context: Walker’s Marginal Efficiency was basically the same as Davidson’s and lower than Woolfolk’s. That’s because the fullbacks were given a lot of carries in short-yardage, lower-efficiency situations. Walker was more efficient, but that was partially based on circumstance.


NCAA Football: Missouri at Arkansas
Albert Okwuegbunam
Nelson Chenault-USA TODAY Sports

Mizzou receivers: Marginal Efficiency

Category J. Moore (WR) J. Johnson (SLOT) E. Hall (WR) A. Okwuegbunam (TE)
Category J. Moore (WR) J. Johnson (SLOT) E. Hall (WR) A. Okwuegbunam (TE)
Catches 65 41 33 29
Yards Per Catch 16.6 17.7 24.8 14.3
Marginal Eff. 17.3% 7.1% 9.6% 23.5%
Marginal Exp. 0.54 0.65 1.37 0.26

Though marginal efficiency and explosiveness can tell us about quality, it’s almost as useful in telling us how a player is used.

Missouri’s offense was almost Army-esque in the way its receivers had roles. J’Mon Moore was the Ahmad Bradshaw of the bunch, the workhorse used in a variety of ways. Emanuel Hall was the home run-hitting slot back. And Albert Okwuegbunam was the red-zone friendly fullback.

We understand the difference between how slot backs, outside receivers, tight ends, etc., can be used in the passing game, and these marginal stats are pretty good statistical methods for defining roles.

National Marginal Efficiency averages

  • Wide receivers: 5.9 percent Marginal Efficiency, 0.17 Marginal Explosiveness
  • Tight ends: 6.7 percent Marginal Efficiency, 0.14 Marginal Explosiveness
  • Running backs: 1.3 percent Marginal Efficiency, minus-0.06 Marginal Explosiveness

Tight ends are used more efficiently than wideouts, but with less explosiveness. (I honestly expected a larger explosiveness difference there.) Meanwhile, running backs are mostly used as break-even options.

Missouri’s differences were more extreme than the average, with Hall’s absurd explosiveness and Okwuegbunam’s excellence in the red zone.

Defensive line

NCAA Football: Texas Tech at Oklahoma
Ogbonnia Okoronwko
Kevin Jairaj-USA TODAY Sports

Oklahoma DL: Marginal Efficiency

Category O. Okoronkwo (DE) M. Overton (DT) D. Ward (DE) N. Gallimore (DT)
Category O. Okoronkwo (DE) M. Overton (DT) D. Ward (DE) N. Gallimore (DT)
Tackles 61.5 31.5 29.5 20.5
TFLs 17.5 3 7.5 1.5
Marginal Eff. -15.0% -24.8% -30.7% -4.8%
Marginal Exp. -0.37 -0.66 -0.49 -0.38

You didn’t really need stats of any kind to tell you that Obo Okoronkwo was Oklahoma’s best defensive player. And if you did, the basic stats did just fine: he had more than double the TFLs (17.5) of anyone else on the team (second-highest: 7.5), and he was the Sooners’ sacks leader by three. He was a one-man havoc machine for a defense that desperately needed one; in fact, he made 17 percent of OU’s havoc plays (TFLs, forced fumbles, INTs, and pass breakups).

Comparing Okoronkwo to fellow OU linemen, however, adds another layer. His Marginal Efficiency (minus-15 percent) was only half as good as fellow end DJ Ward’s (minus-31 percent). How could that be?

The answer lies in the sheer volume of tackles Okoronkwo made from the end position. He was third on the team in tackles, so while he made more havoc plays than anyone else, he also made a lot of tackles downfield, after an opponent’s play qualified as successful.

Expected efficiency and explosiveness differ based on where you line up. Average Marginal Efficiency for a defensive end is around minus-17 percent, while nose tackles (many of which are likely to play in a 3-4 defense) are around minus-19 percent, and those listed merely as defensive tackles are around minus-22 percent. Marginal Explosiveness for each position: around minus-0.44.


NCAA Football: Iowa at Nebraska
Josey Jewell
Reese Strickland-USA TODAY Sports

Iowa LBs: Marginal Efficiency

Category J. Jewell (MLB) B. Bower (WLB) B. Niemann (OLB)
Category J. Jewell (MLB) B. Bower (WLB) B. Niemann (OLB)
Tackles 102.5 65 62
TFLs 13.5 3.5 6
Marginal Eff. -8.9% -2.5% -7.9%
Marginal Exp. -0.51 -0.48 -0.46

Marginal defensive stats go even further down the “how are you using guys?” road, and the linebacking corps is a fascinating example.

Middle linebackers like Josey Jewell tended to have around a Marginal Efficiency rate around 5 percentage points lower (i.e. better) than weakside linebackers and 10 percentage points lower than strong side linebackers. Since we’re seeing a lot more SLBs used as nickel backs (or some approximate equivalent), that would make sense.

Defensive backs

NCAA Football: Sugar Bowl-Alabama vs Clemson
Tony Brown
Kevin Jairaj-USA TODAY Sports

Alabama DBs: Marginal Efficiency

Category R. Harrison (S) M. Fitzpatrick (S) H. Jones (S) L. Wallace (CB) A. Averett (CB) T. Brown (NB)
Category R. Harrison (S) M. Fitzpatrick (S) H. Jones (S) L. Wallace (CB) A. Averett (CB) T. Brown (NB)
Tackles 58.5 49 42 39.5 39.5 24.5
TFLs 4.5 8 1 4.5 4 1
INT + PBU 7 9 4 18 9 3
Marginal Eff. 16.5% -1.4% 30.2% 26.7% 21.5% 2.3%
Marginal Exp. -0.16 -0.24 -0.08 -0.11 -0.01 -0.19

With the increasing array of responsibilities in a secondary, something like Marginal Efficiency and explosiveness can be very useful.

Since passes defensed aren’t yet a piece of this Marginal Efficiency concept, one would expect cornerbacks to have some of the worst numbers, and among Alabama’s, Levi Wallace’s (plus-27 percent) were just that. But the distribution of marginal efficiencies across the safety position — Ronnie Harrison was at 16.5 percent (the average safety is around 20 percent), and Hootie Jones was a whopping 30.2 percent, but Minkah Fitzpatrick’s (minus-1.4 percent) was that of a linebacker’s, as was nickel back Tony Brown’s.

In all, the profile for each level of a defense is as follows:

  • Defensive linemen: minus-19 percent Marginal Efficiency
  • Linebackers: minus-3 percent Marginal Efficiency
  • Defensive backs: plus-23.3 percent Marginal Efficiency

Data like this, however, can tell you not only who was making plays for your defense, but where they were making them and what plays they were preventing.

Next up:

Incorporating these in the arsenal of stats for my 2018 college football team previews — which, incredibly, begin in early February.