This is a guest post from Frank DuPont, author of the book “Game Plan: A Radical Approach to Decision Making in the NFL” which is available now on Amazon. Frank writes about football (real and fantasy) on his site FantasyDouche.com. He also rants about Buffalo Wild Wings, the Bachelorette, the Sylvestor Stallone classic “Over the Top”, and GNR’s video masterpiece “November Rain” on his twitter acccount @FantasyDouche.
By the way, this is a hell of a post from Frank. He does this sort of work on a consistent basis, and I highly, highly recommend his book. – Jonathan
One of the most common themes that appear in emails that I get from fantasy football players is the idea that going into 2012 you have to have one of the top quarterbacks. Usually the email goes something like this: “Running backs can get injured or are otherwise now unreliable. That’s why I think you should go with a sure thing like Aaron Rodgers or Tom Brady early in the draft.” Most of that sentiment I think comes from the fact that last year’s top four or five quarterbacks (Aaron Rodgers, Tom Brady, Drew Brees, Matthew Stafford, Cam Newton) were separated from the rest of the pack.
But the problem with that line of reasoning is that it utilizes only the very recent past to forecast the future. That reasoning is infected with recency bias. I’m not saying that the events of the recent past have no bearing on what will happen in the future, but those recent events tend to have less meaning for forecasting the future than we might tend to think. When those recent events are record-breaking seasons, or what you might call “outliers,” then they probably have even less importance for forecasting the future. The easiest way to get this idea across is with examples.
In 2004, Peyton Manning set the then single-season passing touchdown record in essentially 15 games when he threw 49 touchdowns. Manning was drafted third overall in fantasy leagues the next year and yet he threw only 28 touchdowns (which is less than 60 percent of the 2004 total).
In 2005, Shaun Alexander set the then single-season record for rushing touchdowns when he had 27. Alexander was drafted third overall in fantasy leagues the next year, when he played in only 10 games and had just seven touchdowns.
In 2006, LaDanian Tomlinson broke Alexander’s rushing touchdown record when he scored 28 times on rushing plays. The following seaso,n he would actually lead the NFL in touchdowns again, but did it with just 15 scores (less than 60 percent of the prior year total).
In 2007, Tom Brady broke Manning’s passing touchdown record when he threw 50 touchdowns. The following season he threw zero touchdowns.
In 2007, Randy Moss broke the single-season receiving touchdown mark when he caught 23 touchdowns. During the following season, he caught just 11 touchdowns (less than 50 percent of the prior year total).
The problem with really outstanding statistical seasons is that they often happen during seasons when a lot of things went right. In football, everything goes right just a small percentage of the time. So when I look at the top of this year’s quarterback pool for fantasy leagues, I see a group of guys who are coming off of seasons where a lot of things went right (at least for fantasy football purposes).
But I think an important question is whether the players in that top group of quarterbacks are actually better quarterbacks than Tony Romo. First, Romo’s Adjusted Yards/Attempt is actually higher than Matthew Stafford. So it’s not like Stafford is actually a better quarterback than Romo. He just had more attempts in 2011.
Aaron Rodgers is a ridiculous QB and he had a ridiculous 2011. But is Rodgers destined to throw 45 touchdowns every year? His touchdown numbers from 2008-2010 were 28, 30, and 28, respectively. If you take out an injury-shortened 2010, then Romo’s touchdown totals were 36, 26, and 31. My only point here is that the idea that the top-tier quarterbacks are somehow pre-ordained to put up better numbers is probably off base.
Another way to illustrate how close I think Romo is to the top tier of quarterbacks is by looking at his 2011 season relative to Brady’s. What would Brady’s 2011 season have looked like if he had played a good part of the season with a rib injury, also lost Wes Welker for six games, and then Rob Gronkowski had played injured for most of the season? It wouldn’t have been as impressive would it? We can make a reasonable assumption that it wouldn’t have been as good because Brady did play with a banged up Gronkowski in the Super Bowl and he threw just 276 yards and two touchdowns. You know who averaged about 276 yards and two touchdowns per game last year? Tony Romo. He played most of the season with rib injuries, lost Miles Austin for six games, and had to make do with a banged up Dez Bryant for most of the season.
An interesting exercise I’ve done to see what Romo might have looked like with a healthy Austin and healthy Dez Bryant was to go back and look at quarters where both of the receivers were targeted in the passing game. They didn’t even have to catch the ball. I just looked for quarters in which each receiver was targeted at least once. I found that if you took Romo’s stats during those quarters and multiplied by four to get a full game, Romo would have averaged 310 passing yards per game and about 2.15 touchdowns per game. That pace is within about two fantasy points per game of what Brady averaged last year.
That’s obviously a cherry-picked set of observations meant to make Romo look better, but it’s not crazy to do stuff like that if for no other reason than to give our brains some evidence to counterbalance the powerful effect that the recent past has.
If you asked me who will have a better season this year, Romo or Brady, I would choose Brady and it wouldn’t take me long to decide. But if you give me odds, I might make a different decision. That’s essentially the decision you have this year when you draft your fantasy team. You can choose between Romo and Brady, and if you choose Romo you get superior odds because he’s cheaper. Auction values are a great illustration of cost and right now Brady is going for about $36 in auction drafts while Romo is going for just $23. That’s like walking up to a window in Vegas and betting Romo to have a better season than Brady, and all you have to do is wager $65 to win $100. I’m probably going to be making that bet with a number of my fantasy teams this year. Here’s why:
A large part of my comfort with Romo comes from the presence of Austin and Bryant. Austin has already shown he’s capable of being a number one receiver by putting up a 1300-yard, 11 touchdown season in 2009. At 215 pounds and with a sub 4.5 40 yard dash, Austin fits the profile of a number one receiver as well. Receivers of that size and speed tend to stay relevant all over the field. They don’t run into the problems that smaller receivers do where they are good until their team gets into the red zone.
But most of my Romo enthusiasm really originates in relation to Bryant, who has been disappointing to some. But if he’s been a disappointment, he has been the most impressive disappointing player I can remember.
I have a statistical model that ranks wide receiver prospects coming out of college. That model regards Bryant as a top-level prospect, only behind guys like Larry Fitzgerald and Calvin Johnson on the “can’t miss” scale. Bryant’s 2008 season at Oklahoma State looks like almost no other college season does. He caught 19 touchdowns, but Oklahoma State only threw 25 touchdowns that year. He caught 19 out of 25 touchdowns! He also caught about 50 percent of the team’s receiving yards. Opposing defenses knew that if the ball went in the air, it was probably going to Bryant, and they couldn’t stop him. My wide receiver prospect model gives a lot of credit when receivers catch a large share of their college team’s yards and touchdowns.
While many regard Bryant as somewhat disappointing as a pro, consider that in 2011 at the age of just 23, he was tied for sixth in the league in receiving touchdowns. I have a tough time calling a 23 year old receiver a disappointment when he finishes in the top 10 in the league in receiving touchdowns.
On a per target basis, Bryant has been ridiculously efficient in his first two years. I have a metric that I call Fantasy Points Over Par (with apologies to the Wages of Wins network for hijacking the name “Points Over Par”), which essentially measures how many fantasy points a receiver scores when compared to an average target from that yard line on the field. The whole thing starts with the following graph which shows the expected fantasy points for a pass play based on the line of scrimmage. The line of scrimmage is on the x-axis (YFOG=Yards from Own Goal) while the expected points are on the y-axis.
Expected Fantasy Points/Target Based on Field Position (League Average)
Using the trend line shown in the graph for expected points, I can look at every target for every receiver and calculate whether they scored more or less points than average. To give you some sense as to what very good receivers might look like in terms of Fantasy Points Over Par, the below graphs show Andre Johnson, Calvin Johnson, Dez Bryant and Larry Fitzgerald in terms of FPOP/target for their careers.
You can see that Andre Johnson wasn’t actually much above par until Matt Schaub got to Houston. You can also see that Larry Fitzgerald’s 2010 season was actually below par. It’s also obvious that Bryant is every bit as efficient as these other great receivers were early in their careers. Remember that this is essentially a measure that is blind to opportunity. A receiver only ends up in positive territory by doing more with each target than is expected based on field position. So Bryant’s early career touchdown numbers are actually very impressive.
But it’s also the case that Bryant is probably only now entering the prime of his career. The graph below shows the average percent of peak fantasy production that wide receivers see at each age. These are just averages to illustrate the basic shape of a wide receiver’s career. At 24 years old, Bryant is entering the prime of his career this year.
Let’s get back to Romo now. According to MyFantasyLeague.com, the top five quarterbacks are all going off of the board within the first 20 picks in fantasy drafts. That group includes Aaron Rodgers, Tom Brady, Cam Newton, Drew Brees, and Matthew Stafford. That’s the end of the top tier of quarterbacks.
The next tier of passers are all going between about pick 50 and pick 65 or so. That group contains names like Michael Vick, Tony Romo, Eli Manning, and Matt Ryan. It’s actually worth waiting for Romo in that second tier of quarterbacks because the difference that drafters are assigning to the two groups probably isn’t as large as they think. Romo’s 2011 season was marred by some bad luck that is possible for any of the top passers. Do I know what that bad luck might look like in 2012? No. I only know that in the past, players coming off of huge statistical seasons have had a tough time reproducing that result the next year. It’s not impossible, it’s just not expected.
So rather than assign a premium to the top tier of quarterbacks based on what they did the previous season, I would rather draft a quarterback like Romo who is going to have a lower average cost at his position. Then I’ll hope that some amount of bad luck that he had the prior season doesn’t manifest itself in 2012. If he has that same amount of bad luck, then all I did was pay for what I got, which is a quarterback who will put up middle of the road numbers. But it’s not unrealistic at all to think that if Romo and his receivers stay healthy, he could put up a season closer to what Brady and Stafford put up in 2011. Is it guaranteed? No. But because of Romo’s reduced cost compared to the top tier, you’re getting what I think are really good odds to go that route.