The Football Outsiders Almanac for the 2009 season was recently released, and the author of the Lions' section of the book, Bill Barnwell, answered five of my questions about last year's team and what to expect out of the '09 Lions.
1. Based on a couple different formulas used in Football Outsiders, last year's 0-16 Lions weren't actually the worst team of all time, statistically speaking. Obviously they will always be remembered for going winless, but how much better than their record were the '08 Lions?
Bill Barnwell: Well, it depends on how you calculate it. Their points scored and points against were indicative of a 3-13 team. DVOA, our stat which measures performance versus the league average on every play after adjusting for down, distance, situation, and opponent, would consider them to have played at the level of a 2-14 team.
DVOA suggests that the Lions were the worst team in the league, but only by a slim margin over the Rams. It also pegs the Lions to be a garden-variety really bad team as opposed to the worst team in NFL history, which is an "honor" we might bestow on a team like the 2005 49ers.
2. In your Lions preview, you mentioned how Jim Schwartz is really into statistics and the type of things you find in Football Outsiders, so much so that he consulted with you guys before and once contributed to the book. What was that project all about?
BB: Well, Jim didn't directly contribute to the book. We've worked with several NFL teams on projects, including Jim while he was in Tennessee. As you might suspect, we can't talk too openly about what we've worked on and for whom.
3. Lions fans know how good Calvin Johnson is, but explain DVOA and how it showed that he was on another level compared to his teammates last season.
BB: We mentioned DVOA above; we apply the same concept to every play that a player was involved in, so we can measure how a receiver did with the passes he was thrown.
Here's each receiver with more than 10 targets' DVOA:
Calvin Johnson 9.1%
Shaun McDonald -23.3%
Mike Furrey - 16.7%
John Standeford -26.6%
Keary Colbert -25.2%
Michael Gaines -16.1%
Casey FitzSimmons -44.1%
John Owens -20.7%
You'll note that Johnson is the only receiver even close to above-average, let alone actually above it. That also doesn't consider the fact that Johnson was routinely double or even triple-teamed, and was thrown passes out of sheer desperation and panic at times.
You probably didn't need us to tell you, but allow us to do so: Calvin Johnson is really, really good.
4. You guys weren't fans of the Lions drafting Matthew Stafford with the top pick based on his college numbers. What do his numbers say that make you think he won't end up being worth the top overall selection, and how good of an indicator of success in the NFL are college statistics?
BB: Our Dave Lewin found several years ago that games started and completion percentage were extremely reliable in predicting the NFL success of quarterbacks drafted in the first two rounds of the NFL Draft. Those numbers for Stafford aren't particularly good; his list of comparable quarterbacks includes Matt Ryan (good), but also Patrick Ramsay, Jake Plummer, Shaun King, and J.P. Losman. That's not exactly an inspiring list of comparable players.
We're still slowly sifting our way through college statistics to try and find metrics and data that means something at the pro level. The quarterback research conducted by Lewin several years ago is the best thing so far. Truthfully, the biggest project at the moment comes with regards to actually acquiring data; there's no repository of college statistics, so we have to compile it school-by-school and year-by-year.
5. The Lions' mean projection in your preview is 5.8 wins. How was that number calculated, and what are the Lions' best- and worst-case scenarios for 2009?
BB: We have a projection system that takes into account about a hundred variables for each team that have shown some significant level of predicative power on team performance over the past 15 seasons. After punching in all the variables and turning the crank on our computers a bunch, we get a projected DVOA for each team's offense, defense, and special teams.
From there, we simulate the season 10,000 times using that projected DVOA using expected levels of variance, home field advantage, and other assorted variables. We average up the win totals from each of those 10,000 seasons and produce an expected number of wins.
Last year, we had the Lions pegged for 4.3 wins, the third-worst total in the league, so we weren't THAT far off. The only teams we had below them were the Raiders (not bad) and the Falcons (can't win 'em all).
Thanks to Bill for taking the time to do the interview.