Consider this an extension of last week's Florida preview week as it wasn't done in time to be posted then.
In the roster post from out Florida preview week, I speculated that Jeff Demps would end up as Florida's primary running back. It partially had to do with Chris Rainey and Emmanuel Moody being injury prone, but mostly it's because I perceive Demps to be the most effective of the three. Demps has since been named the starter for now.
Well, I've also been following the discussion about running back efficiency on Chris Brown's Smart Football blog. It began by posing the question of whether you want a boom-or-bust back like Barry Sanders or a grinder like Jerome Bettis. The debate continued with some heavy Football Outsiders flavor. Most recently, Advanced NFL Stats weighed in with some studies about the distribution of rushing gains.
There is yet to be a firm consensus, but on thing everyone agrees on is that yards per carry is inadequate for comparing runners. For one thing, it penalizes players for gaining three yards on third and two and rewards them for gaining nine yards on third and ten. It also can be skewed greatly by long runs. YPC has its place, but it's not everything.
Anyway, I decided to piggyback off of the Advanced NFL Stats post and look at rushing distributions. It struck me that AFS found that just about every NFL runner has a median carry of three yards and the distribution of yardage on carries for any given back is virtually identical to that of the entire league. That means the two things that separate the exceptional backs are avoiding losses and big gains.
I looked at Florida's top four rushers from last season (Tim Tebow, Rainey, Demps, and Percy Harvin) to see who was most effective and how this year's running game should be structured. I did not include Emmanuel Moody's stats because almost all of his carries in 2008 came in garbage time. It also should be noted that Rainey has gained weight at the coaches' request this off season, so he will be a bit different as a back this year.
Here's a quick reference chart of the important stats to give you a glance at how the Gator rushing attack worked in 2008. "Team" here refers to just the four listed players combined. In order to get the distributions, I had to go through the play-by-play accounts of the games, and it would've taken a long time to include every single rush. Somehow I ended up one carry short for Harvin, but I've been through the play-by-plays twice now and I'm not doing it again.
Also, Tebow's numbers do not include sacks, and all stats against the Citadel (I-AA) have been thrown out.
| Category | Team | Tebow | Rainey | Demps | Harvin |
|---|---|---|---|---|---|
| Carries | 379 | 159 | 77 | 77 | 66 |
| YPC | 6.46 | 4.58 | 6.81 | 7.71 | 9.12 |
| Median | 4 | 4 | 4 | 4 | 5 |
| Pct. are Gains | 83.4% | 84.9% | 77.9% | 81.8% | 87.9% |
| Gains 30+ yds | 14 | 0 | 4 | 7 | 3 |
| Pct. of total | 3.7% | 0.0% | 5.2% | 9.1% | 4.5% |
| YPC, no 30+ gains | 4.82 | 4.58 | 4.38 | 4.10 | 6.75 |
Demps and Rainey have nice YPC numbers, but Harvin takes the cake. In fact, Harvin was the best in every category except gains of 30+ yards as a percentage of total carries. His yards per carry is the highest, his median is highest, he did the best at avoiding losses or no gain plays, and his yards per carry with long runs removed is head and shoulders above the rest.
What is a bit surprising is that net of long runs, Tebow has the next highest YPC. He gets a reputation of grinding out short runs while Demps and Rainey go for longer, but if you ignore a combined 11 total carries from those two, they aren't as effective a down-to-down runner as Tebow is.
What is not a surprise is that Rainey has the lowest percentage of gains. He is a shifty guy, but a consequence of his confidence in his ability to make people miss is that he dances around too much and takes losses.
Demps' big play percentage is truly remarkable though. He didn't just end up best, but he doubled Harvin's percentage and nearly doubled Rainey's too. That would seem to make him a boom-or-bust back, but he went boom more than Rainey did and is not all that far behind Tebow at avoiding busts.
While recording all the rushes, I included down and distance too. That allowed me to see who was most effective when. I looked both at YPC on each down (capped at 30 to avoid too much skewing) and the percentage of the time that the player picked up a first down. If a player had fewer than five carries on a particular down, that was excluded.
| Player | Down | YPC | Pct. 1st Dn. |
|---|---|---|---|
| Tebow | 1 | 4.27 | 20.34% |
| Tebow | 2 | 4.72 | 29.82% |
| Tebow | 3 | 5.41 | 74.29% |
| Tebow | 4 | 1.50 | 75.00% |
| Rainey | 1 | 6.92 | 30.77% |
| Rainey | 2 | 3.70 | 26.67% |
| Rainey | 3 | 8.71 | 57.14% |
| Demps | 1 | 6.31 | 20.83% |
| Demps | 2 | 8.69 | 46.15% |
| Harvin | 1 | 8.28 | 27.91% |
| Harvin | 2 | 7.11 | 63.16% |
Tebow got better from first to second to third down, despite being called on to convert third and short frequently. The YPC on fourth down looks bad until you consider that every fourth down carry was with one yard to go. Harvin had the best first down average, but Rainey picked up fresh sets of downs more often. The same was true for Demps (higher average) and Harvin (more first downs) on second down.
* * *
Now, there are other factors than just running effectiveness that will determine who gets on the field when. Blocking is a big factor, for one. If you go to Urban Meyer's website (turn your sound off first) you can see a video of a drill that has running backs picking up pass rushers, and Rainey and especially Demps cannot block an edge rusher at all right now. Brandon James can though, and walk on back Chris Scott (who's been pushing for playing time) did a nice job as well. That could mean James and/or Scott will end up as primary guys for draw plays from passing formations since they'd be more likely to be in on pass plays for their blocking.
However, it's pretty clear right now that Demps is the top back on the depth chart, followed by Rainey and then the injury prone Moody. It also should be clear from the past couple years that Tebow will be getting a lot of carries, and it's doubtful that anyone would be surprised to see him lead the team in rushes again.
Florida will definitely miss Harvin in the running game since he was the best ball carrier in just about every way. No one will step in an produce at that level. We haven't seen too much Moody yet, but he's likely to be somewhere between Tebow and Rainey/Demps in these categories.
The running game always begins with Tebow because having a mobile quarterback gives Urban Meyer the "extra blocker" that he covets so greatly. Tebow also gets stopped for no gain or losses less frequently than the other two returning guys from this study, and that's important to a guy who is old school in his heart like Meyer. Rainey might get slightly more on a down-to-down basis, but Demps makes sense as the top back. He loses yardage less frequently than Rainey does and he breaks off big plays at a startlingly high rate.
The Gator running game should be in good shape in 2009 despite Harvin's loss. The rest of the guys aren't exactly chopped liver, and the Gators edged run-heavy Alabama in rushing yards in the SEC title game with Harvin on the sideline. If you want to see the distribution graphs for the four above players versus the team, they are here:
