LenDale White, Chris Johnson, running back use and success

I believe I’ve mentioned on here a couple of times that, in terms of evaluating running backs, I don’t particularly care for yards per attempt. After all, a lot of runs go for about 2 yards, and the difference between average 4.5 and 4.3 yards per carry is 1 gain for 20 instead of 10 yards every other week. Instead, I prefer the statistic success rate, which I learned about from Football Outsiders.
What success rate is is a measurement of how close a carry comes to picking up first down yardage-a carry on first down is a success if it gets 40% of the yardage to gain, on second down carries 60%, and third and fourth down carries 100%. And, if you run the numbers for the Titans this year, you find that, through four games, LenDale White and Chris Johnson, have the exact same Success Rate, 43.3%. Johnson is averaging over 5 yards a carry, White under 2.7. Despite this massive disparity, have they really been about as good this year?
The answer, unsurprisingly, is of course not. (What, you didn’t think I was going to write a post defending White, did you?) The explanation is an old friend, and the reason I prefer Success Rate to yards per carry as a measure-the addition of context. Johnson and White have had roughly the same number of carries (67 and 60, respectively), but when you look at carries by down-and-distance, that Johnson is much better than White becomes clear.
Take, for instance, the most common play in football, 1 and 10. Helpfully, each of White and Johnson has had 30 of his carries come in this situation. 15 times, a Johnson run on 1&10 has gained at least 4 yards-a very good 50% Success Rate. White, by contrast, has had a successful carry only 10 times on 1&10-a poor 33% Success Rate. Put it like this-the Titans are 50% more likely to end up in 2&long if they give White the ball on first down than if Johnson gets the ball.
And 2nd and long is where the disparity between the two backs, and the reason for Johnson’s equal Success Rate despite being a superior back, comes clear. I noted based on preseason stats that the Titans had lots of difficulty running on 2nd and medium and 2nd and long. Those difficulties have continued in the regular season. Overall, White has a 44% Success Rate on 2nd down, while Johnson is at 39%. Looking inside those numbers, though, and the problems of 2&7+ become apparent. White has carried the ball 8 times in such situations, and has been successful 3 times. Johnson has superficially good per carry stats, thanks to his 51 yard run against the Bengals on 2&12, but overall has only been successful on 3 of 16 carries on 2&7+, a 19% Success Rate. Now, I don’t have league-wide stats for rushing success on 2&long, or even Titans stats for prior years, but it’s clear to me that having to run on 2nd and long is a losing proposition, and Johnson’s looking worse because of it.
What about third down, you ask? White has converted both times on 3&1, 1 of 3 on 3&2, and failed on 3&5, for a total of 3 conversions on 6 attempts. Johnson has, by contrast, only converted on 4 of his 11 3rd down carries. Once again, a little context is in order. Johnson has 1 carry on 3&1, 1 on 3&2, and 1 on 3&3, and converted all of those. He also converted 1 of his 3 carries on 3&5, for a total conversion rate of 67% on his 6 carries that are akin to White’s 6 carries. Johnson, though, has 5 additional carries, each from 7 yards or more out. Unsurprisingly, he has not managed to convert on any of these plays. Now, the leaguewide Success Rate on runs on 3&long isn’t 0, but that doesn’t mean it’s actually any good-that Johnson has failed in this situation is not necessarily a negative reflection on his ability.
So, how should we get a true comparison between White and Johnson? The difference in usage is that Johnson has gotten 2/3 of the runs on 2&long and all of the runs on 3&long. So, let’s pretend like he got White’s workload-half-weight his 3/16 on 2&long and ignore his 5 carries on 3&long. Take those away, and what do you have? White’s Success Rate is still 43.3%, while Johnson’s has now improved to 50.9%. What my eyes and the per-carry statistics were telling me is now confirmed-Johnson has clearly been more successful than White. He just didn’t look better because he was put in a position to fail more often, a position he was put in precisely because he’s a better back. When it comes to football and statistics, even in the use of more advanced statistics, context still matters.