Stat Analysis
Advanced analytics on player and team performance

Portis Uber Alles

by Aaron Schatz

I was talking about the upcoming season with a few guys the other day and they asked me who I thought was going to be this year's rushing champion.  I had to think about it for a second.  From my article on Ricky Williams, we know that Ahman Green is likely to be headed for some decline and Jamal Lewis is even more likely to be headed for some decline.  Priest Holmes doesn't really get enough carries, since he's involved in the passing attack so much.  LT would be a good choice, but after two big steps forward it may be time for him to take a step back, and he'll be the first running back south of the 49th parallel to sometimes see 12 men in the box.  Corey Dillon would just be wishful thinking on my part, and Minnesota is too busy splitting carries among its backs.  I said, "Clinton Portis."

Nonetheless, it seems that Clinton Portis is not getting the proper respect this preseason.  Look over at the expert rankings at, and you will find that the consensus ranking of Portis is #5 among running backs.  Not a single expert has him above #3.  Three experts don't even have him in the top eight.  Mark Wimer has him twelfth.  TWELFTH!  Behind Domanick Davis and Kevan Barlow!  I mean, I suppose Portis could be hit by a bus next week but otherwise that is totally absurd.

This Portis non-love is based on two pieces of conventional wisdom: First, that any running back can succeed in the Denver system, and second, that Washington has not been a good rushing team the past two years under Steve Spurrier.  Both of these statements are not quite as true as you think.

It does seem at first glance like any running back can succeed in the Denver system.  Three different running backs have run for 1500 yards in a season for Denver since 1998, and a fourth (Olandis Gary) ran for 1000 yards in only 12 games.  None of these backs were first round picks.

But there is a difference between succeeding in the Denver system, and succeeding like Clinton Portis.  Terrell Davis was more of a workhorse, and I mean no disrespect to Otis Armstrong, but Clinton Portis on a per play basis has been the best running back in Denver Broncos history.

In another article being posted on Football Outsiders today, Michael David Smith talks about comparing a starting running back to his backups as a way to try to determine how much of the back's success is due to his ability and how much is due to his linemen.  To show just how good Portis has been over the past four seasons -- when compared to other running backs playing in the same system and behind (roughly) the same linemen -- here are all the Denver running backs of the past four seasons with at least 20 carries, ranked in order of DVOA.  For you newcomers -- after all, this article is a completely blatant attempt to get linked on Denver and Washington message boards -- that's our innovative statistic that takes every single play in the NFL and compares it to league average based on down, distance, and situation.  (Learn more here; the numbers will be slightly different than what is listed on the running backs page on the website because this is a slightly updated system.)

Player Year DVOA VOA Yds Car Yds/Car
C.Portis 2002 24.5% 31.3% 1508 273 5.52
C.Portis 2003 15.2% 22.4% 1591 290 5.49
T.Davis 2001 12.7% 13.6% 701 167 4.20
M.Anderson 2000 9.3% 15.1% 1500 297 5.05
O.Gary 2001 2.6% 4.8% 228 57 4.00
T.Davis 2000 -0.2% -5.0% 282 78 3.62
M.Anderson 2002 -2.0% 3.1% 386 84 4.60
O.Gary 2002 -4.3% -7.6% 147 37 3.97
M.Anderson 2001 -11.0% -11.7% 678 175 3.87
K.Coleman 2000 -27.2% -31.2% 189 54 3.50
M.Anderson 2003 -29.6% -21.8% 257 70 3.67
Q.Griffin 2003 -32.4% -28.5% 353 94 3.76

Unfortunately, I cannot go back with DVOA before 2000 -- and that includes Olandis Gary's one year of glory -- but I can go back with yards per carry. Olandis Gary in 1999 averaged 4.20 yards per carry, more than a yard per carry lower than Portis in each of the past two seasons. Terrell Davis in his four years in Denver never averaged as many yards per carry as Clinton Portis. Even in the year he topped 2000 yards, he averaged 5.12 yards per carry.

Now, after a moment of silence for the career of KaRon Coleman, notice who is at the bottom of that list.  Yes, that's Quentin Griffin, your new starting Denver running back.  Based on his performance in 2003, there is little to indicate that Quentin Griffin will be able to replace Clinton Portis as the lead back for the Denver Broncos. Griffin averaged only 3.4 yards per carry on first down, compared to 5.4 yards per carry for Portis. Griffin averaged only 4.2 yards per carry on second down, compared to 4.6 yards per carry for Portis. Griffin's numbers look even worse when you take out the blowout win over Indianapolis, in which he gained 136 yards against a terrible Colts run defense. Without that game, Griffin averaged only 3.3 yards per carry for the season.

There is nothing to indicate that Griffin will be able to step up his game to replace Portis in the Denver offense.  There isn't really much to indicate that he should be the man given the opportunity.  The one thing that he has going for him is that 94 carries is not enough to make a definitive statement about your ability.  But for him to go from a rookie year like this to anywhere near the level established but Portis or Terrell Davis would be almost unprecedented.

If Griffin proves to be a failure after a couple of games, who gets the ball next?  Well, the Broncos signed Garrison Hearst to be the grizzled veteran of the running game, but the sun is clearly setting on his career.  In each of the last two years, Hearst has had less carries, less yards, less yards per carry, and a lower DVOA than the year before, and there is no reason to see this trend reversing in 2004.  Mike Anderson is still around, but his history shows that you can't be sure what you'll get out of him, and Mike Shanahan seems reticent to just give him the ball and let him play halfback again.

This is where I was supposed to talk about my belief that Tatum Bell was destined to be the Denver starting running back by midseason.  After all, he is the highest-picked running back in the history of Shanahan's tenure, ten picks higher than Portis in the second round.  Then, Bell held out for a few days, finally signed, reported to camp, and broke his finger, meaning he's now going to miss another few days of camp.  Rookies who miss large parts of their first training camp don't really have a good track record, and this can't be giving Mike Shanahan happy Tatum Bell thoughts.  Of course, many a fantasy football owner has seen their season wrecked by an attempt to read Mike Shanahan's mind on the issue of starting running backs, so perhaps this all will end up meaning nothing and Bell will be starting by midseason anyway.

Having frowned upon Denver's replacements for Clinton Portis, let's turn back to Portis himself.  Why is Portis not being considered in the same class as LaDainian Tomlinson and Priest Holmes?  Well, part of it is the idea that he's going from a team with great offensive line to a team with an offensive line that has trouble run blocking.  I'm not really sure that's accurate.

Another new statistic we have created at Football Outsiders is called adjusted line yards.  The goal is to try to separate the performance of the offensive line from the performance of the running backs by only measuring the offensive line on the first ten yards of each run.  The idea is that once a running back gets past a certain point, success is based more on his ability to juke and break tackles and less on the blocking scheme that created his holes.

For 2003, Washington actually ranks tenth in the NFL in adjusted line yards.  Denver is behind them, ranked #13.  That rating includes all the Denver backs, Portis as well as Griffin and Anderson, so it gets dragged down a bit by the poor performances of Portis' backups.  But it is still strange to think that Washington and Denver might have been equally good at clearing out room for their running backs last year.

If that's true, though, where does the difference come from when you compare Portis' amazing numbers and the mediocre stats of last year's three-headed Washington rushing attack?  Well, if we are going to measure runs by cutting them off at ten yards, we can also measure how much of a team's total running game came from big gains once the back had broken free.  You'll find that stat under the column marked "10+ Yards."  Washington, with 11% of rushing yards coming after backs had already gone ten yards, is ranked #28.  Denver, with 25% of rushing yards coming after backs had already gone ten yards, is ranked #5.  Check out 2002 and you'll see the ranks are #24 for Washington and #10 for Denver.

All these numbers tell a story.  Washington's offensive line can clear out holes for the running backs to gain good yardage.  Once Betts and Canidate were in the open field, however, they weren't very good at avoiding defenders.  Portis, on the other hand, picked up a huge amount of his yardage on those long runs where he was on his own, with no offensive linemen nearby.  Perhaps part of that is the renowned blocking ability of the Denver wideouts, but for the most part, that's Portis being one heck of a back.

(Egads, did I just channel John Madden there?  Allow me a second to smack myself.)

What happens when you stick the running back who is very good in the open field with the offensive line that is underrated when it comes to blocking for mediocre backs, then add an old-school coach whose teams have finished in the top five in rushing attempts seven times, including a year where the team was only 10-6 and was actually outscored over the course of the year?  Well, if Portis manages to play all 16 games for the first time in his career, I think you get the NFL rushing champion for 2004.

That means I think Denver made a bad trade, right?  Actually, no.  Even though I think the Denver rushing attack is going to be worse than people think this year, I do not think they made a bad trade.  I've already expounded upon what my stats have to say about Champ Bailey; this article discusses how Washington was one of the best teams last year at defending against the other team's top wideout, and was the worst team playing pass defense against everyone else.  If the trade was Portis for Bailey, I would consider it a win for Washington.  But it wasn't.  It was Portis for Bailey and the pick that became Tatum Bell.  There's a tiny chance that Bell will develop into Portis; there's a bigger chance that Bell will develop into, say, Deuce McAllister or Travis Henry.  If Bell can develop, or if Griffin can make a massive unexpected improvement this year, Denver took a moderate downgrade at running back and got a big upgrade at cornerback out of it.  The team as a whole should be as good, or better, than last year.  But let's not all go kidding ourselves into thinking that they won't miss Portis.


There are no comments yet.