Statistical Deep Dive: FSU’s Cook vs. Clemson’s Etienne

Florida State Seminole football fans are no stranger to being slighted by ESPN. While much of hashtag FSU Twitter spends their days defending the honor of the ‘Noles, usually it is disputing the anecdotes of bloviating pundits, poorly-coiffed tennis aficionados and Chris Fowler.

So the statistical comparison from ESPN’s highly respected David Hale that concluded Travis Etienne is a better college football running back than Dalvin Cook took many aback. It was not long before Hale’s twitter feed was met with criticism, both with anecdotes and with statistics.

Even Miami Hurricanes blogs got in on the action, siding with the ‘Noles and Cook:

Let’s start at the top of Hale’s statistical rundown:

This seems fair, on its face. Hale assembled several basic statistics, and it’s hard to disagree with the base argument: Travis Etienne had a statistically better career than Dalvin Cook when you look at yards/rush, career touchdowns, etc.

At Tomahawk Nation, this piqued our curiosity. We focus on advanced metrics when it comes to statistical analysis, because the rote statistics of the past can leave much to be desired. Our founder emeritus Bud Elliott would chastise us for using per game or per season stats rather than per play stats, and we go even deeper than that - now that more advanced metrics are available to us (through TN’s subscriptions to multiple statistical services as well as increased public availability).

Hale’s statistics do not isolate for other factors in play. The most obvious of these would be the poor FSU offensive lines of the era, as well as the poor quarterback play in Cook’s final two years - and, frankly, middling play of Jameis Winston’s 2014 first halves in his debut campaign. While Etienne was similarly hampered by Kelly Bryant as a freshman and a poor offensive line as a true senior, three years of future #1 overall pick Trevor Lawrence does matter immensely.

Hale rightfully said that this was anecdotal and/or hard to factor in - which it is, unless you dive in deeper.

Let’s take a look at several metrics from the two back’s college careers:

A quick rundown on definitions here can be found in Football Outsiders’ Offensive Line Stats page. In addition to FO’s two OL metrics, I did a simple average of these to one number for convenience sake. Let’s call this the Juan Line Average Metric (JLAM). Additionally included are RB YPC - running back yards per carry, OYPP - offensive yards per play for the season overall, OYPA - offensive yards per pass attempt for the season overall, OFEI - the offense’s overall FEI ranking, and Passing SP+ rating - the SP+ rating of play-to-play efficiency of the pass game.

Key note: 2019 and 2020 Pass SP+ is not readily available, as ESPN does not offer these metrics out as easily dispensable as Football Outsiders did for previous seasons. That said, I think it is very fair to assume that Clemson with Lawrence and Tee Higgins, Justin Ross and co. were a top 15 passing offense or better those two years.

Dalvin Cook played behind a bottom half FBS offensive line his first two seasons, and a top 20 unit his final season. Travis Etienne played behind top 25-30 units his first two seasons, an elite unit his junior year, and a bottom half Power five (still upper half of FBS) unit his senior year.

This is where Hale’s use of old school statistics missed the mark:

Yards after contact and NFL draft numbers were the best metric available seven or eight years ago. Fortunately, Brian Fremeau, Bill Connelly, Football Outsiders and a bevy of other pioneers brought us out of those dark ages.

Sure, Dalvin Cook has less yards after contact than Etienne. But his yards before contact were 12% better, despite never playing behind a better unit. Opportunity rate tells us if an offensive line gives running backs an opportunity. Cook had far worse opportunity rates than Etienne: Cook behind 79, 54 and 27 averaged 3.54 yards prior to contact and Etienne behind 26, 33, 1 and 66 averaged 3.11 yards prior to contact.

Despite never having an offensive line in the top 25 of opportunity rate, Cook avoided contact for nearly a half yard more than Etienne - per rush.

So the offensive line argument actually goes significantly in Cook’s favor. Cook did more with much less.

Hale is correct here, in my opinion. Etienne was marvelous for Clemson his freshman year, despite playing with Kelly Bryant. Going back to my chart above, they notched a top ten offense despite a 61st ranked Passing unit and a good, but not great, top-30 offensive line.

You can probably most accurately compare this season to Cook’s sophomore 2015 campaign: the ‘Noles had a 45th ranked passing offense, a bottom half Power 5 run blocking offensive line, and managed the 28th ranked offensive unit.

In those above seasons, Etienne averaged 7.16 yards per rush, and Cook averaged 7.38. The difference is nominal, frankly, and saying either was better is really just splitting hairs. Both were fantastic in their only seasons playing without generational QB talent.

Next, let’s compare their performances behind bad offensive lines. I think this is where Cook really shines - and where he is shining for the Vikings in the NFL (and on your fantasy team).

Cook had his best season behind what can best be described as a mediocre unit in 2015: 54th in opportunity rate, 107th in stuff rate, and Cook managed to put up that 7.38 yards per rush. This was with the 46th best passing offense.

Etienne faltered drastically when he did not have a top 30 OL unit in his senior year. The 2020 Clemson offensive line was 66th in opportunity rate and 36th in stuff rate, and Etienne dropped to 5.44 yards per rush, over a yard and a half below his career average. This was with the 11th ranked overall offense, and Trevor Lawrence’s career best yards per attempt.

There is no one single statistic, rote or advanced, that can definitively say Dalvin Cook was a better college running back than Travis Etienne, or vice versa.

Etienne and Cook will go down among the best running backs in ACC history. Both are excellent, and there is no statistical argument that one is better than the other.

That said, I take Cook every day and twice on Sunday.