The football column for Week Nine of the NFL season is going to focus on two of the game’s elite players in 2007 – Derek Anderson and Adrian Peterson. Of course, before we get to the stories, let’s start with the Week Nine rankings.
The Surprising Derek Anderson
Here are the top five quarterbacks thus far in 2007 (in terms of QB Score per play): Tom Brady, Tony Romo, Peyton Manning, Derek Anderson, and Jeff Garcia. To see Brady and Manning on this list is not too surprising (although, as noted last week, Brady is not typically this good). Romo was a top five quarterback last year, and was recently rewarded with a $67.5 million contract. Even Jeff Garcia is a veteran of past Pro Bowls. But Derek Anderson?
Anderson finished the 2006 season with a 0.91 QB Score per play, a mark that is well below average (average QB Score per play is 1.85). Such a poor performance (and you didn’t need to see QB Score to know Anderson didn’t play well), led the Cleveland Browns to surrender a 2008 first round pick (and a 2007 second round pick) to take quarterback Brady Quinn in the first round of the 2007 NFL draft.
Why did the Browns go to such lengths to get Quinn? Again, an above average QB Score per play is 1.85. The last time a Browns quarterback attempted 224 passes in a season (the minimum necessary to qualify for the NFL’s quarterback rankings) and surpassed the average QB Score was in 1995 (Vinny Testaverde). Yes, since the Browns returned to Cleveland, this team has not had one above average quarterback. Hence you can see why this team would give up so much for Quinn.
After eight games this year, though, the Browns not only have an above average quarterback but one of the top five signal callers in the game. And this top five quarterback was on the roster last year. So although the history of the Browns suggests quarterback was a problem area, shouldn’t the coaches have known that Anderson was a future star? Shouldn’t this team have kept their 2008 first round pick and passed on Quinn?
Thinking about Basketball For a Moment
Let’s think about these questions this way. LeBron James was taken by the Cleveland Cavaliers in 2003. In 2003-04, LeBron was only an average player. Since then, though, he has been consistently a top five small forward. In 2004 the Cavaliers could have traded up in the draft to get Luol Deng. With LeBron on the roster, though, such a trade would have been questioned. Why trade up for a great small forward when you already have one that should be great on the roster?
Now in basketball, it might make sense to acquire a second similar player because one player doesn’t play the entire game. And you can also switch a small forward to another position like shooting guard. In football you do not have this flexibility. Teams do not typically split playing time at quarterback. Furthermore, no one suspects Quinn was drafted to play another position. So again we ask, why take Quinn?
Answering the Question
The answer is quite simple. Obviously the Browns didn’t know that Anderson would be this good.
But why didn’t they know this? The Browns had seen Anderson in practice and games for an entire year. They should have known what he could do. Yet, it’s clear, what he is doing this year has caught the Browns decision-makers by surprise.
Hence we come to the point of today’s column. Decision-makers in football have a tremendously difficult time predicting future performance.
In the NBA we can see teams have two specific problems. The impact of inefficient scorers tends to be overvalued, while prolific rebounders are undervalued. The most famous example of the former is Allen Iverson, whose departure from Philadelphia actually improved the team in 2006-07. But for players like Kevin Garnett or Ray Allen (two efficient scorers), people expect their addition to be helpful. And the early returns seen in Boston confirm that expectation.
In football, there may or may not be a problem with how the data is interpreted. But in the end, it also doesn’t make much difference. Even if you had the best model possible, past performance in the NFL doesn’t predict the future very well. Consequently Brady Quinn decisions are difficult to avoid. The Browns looked at Anderson last year and concluded that he wasn’t a franchise quarterback. And it may still be the case they are right. After all, we are only eight games into the 2007 season. But it’s also possible that Anderson is a legitimate elite signal caller. If that turns out to be true, the drafting of Quinn will prove to be a mistake. Unfortunately, it’s a mistake that probably couldn’t be avoided since past performance in the NFL cannot accurately predict the future.
The Greatness of Adrian Peterson
After eight games Adrian Peterson has rushed for 1,036 yards. Which puts him on pace to rush for over 2,000 yards. This is pretty good, considering that Peterson hasn’t even started much of the season.
Peterson is also averaging 129.5 yards per game, a mark that has not been matched since 1994 (when my data set – which I took from Yahoo.com – began to track fumbles lost). When we turn to RB Score per game – which considers rushing, receiving, and fumbles – we see that only twice has a player bested Peterson’s mark of 80.3. And both times this happened the running back was named Marshall Faulk.
Faulk’s RB Score per game was 84.8 in 2000 and 84.3 in 1999. Every other running back who has played since 1994 has failed to hit the 80 threshold. In fact, only five running backs – Priest Holmes (2002), Barry Sanders (1997), Marshall Faulk (2001), LaDainian Tomlinson (2003), and Tiki Barber (2005) – has posted a per game mark greater than 70. And only seven more backs – Clinton Portis (2003), Charlie Garner (2002), LaDainian Tomlinson (2006), Barry Sanders (1994), Garrison Hearst (1998), Steven Jackson (2006), and Brian Westbrook – got past the 60 mark. And only 17 more backs (and no, I am not going to name all of these) got past the 50 per game level. In sum, of the 553 observations we have of a running back rushing 100 times in a season since 1994, only 31 (or 5.6%) had an RB Score per game greater than 50.
This year there are four running backs who have passed the 50 mark. These include Peterson, as well as Ronnie Brown (73.9), Westbrook (73.0), and Joseph Addai (56.6). Although these backs are playing well, they should take note of the other running backs I have mentioned who are still playing in 2006. Tomlinson in 2007 has only posted a per game mark of 44.4. Yes, this is good, but not quite what he did in 2003 or 2006. Steven Jackson (whose 77 carries is a bit short of my threshold of 10 attempts per game I have arbitrarily chosen for my weekly rankings) only has an RB Score per game of 4.8. And Portis only has an RB Score per game of 26.9 this season.
The decline we observe with respect to these players should be noted by Peterson, Brown, Westbrook, and Addai. Today each player is at the top of the league. But soon – in the next game or next season – we might see quite a different result. And as we saw with respect to Derek Anderson, it’s not likely we can predict when this will happen.
For more on QB Score, RB Score and what these metrics mean see