I finished up the comparison of Fournette vs. Guice that I mentioned. All of the raw data is on a google doc here: LINK Apologies in advance for any readability issues.

Methodology is fairly simple: Take the carries and rushing yards against each team on the schedule, subtract the stats that Fournette and Guice put up against them, and calculate average yards per carry against each team minus those two. Multiply average yards per carry allowed by the number of carries each back had against that team that season. That gives the expected rushing yards in each game given the number of carries each back had. Total up the expected rushing yards, compare to the actual yardage total for each player, and divide by number of carries to see how much better than average each player played. I took out games against Sam Houston State and Jacksonville State, due to problems with comparing statistics for teams at a lower level of play, but I don't think it impacts the final results too much.

Or, to put it more simply, how much better did they do than an average back would have done with the same carries against the same opponents.

Given Fournette's carries and opponents, an average running back would have been expected to put up 2551.7 yards on his 603 carries. LF7 actually totaled 3738. This is an incredible 2 yards per carry above average.

Guice, on 215 carries, has an expected 946.8 yards. He actually put up 1668 yards. As good as Fournette's yards above average was, Guice's is an even more amazing 3.4 yards per carry above average.

There are potential pitfalls with these numbers. For one, Guice has a much smaller sample size than Fournette; that makes a big difference in this kind of an analysis. For another, we have no way of knowing how offensive scheme and defensive attention changed for each player. Given that both backs were playing for the same team in run-heavy schemes, I don't think there's anything too terrible about making this kind of comparison, but I'd hardly consider it gospel truth either.