In this week's Varsity Numbers, Bill Connelly revisits some measures and concepts: Adjusted Scores, Covariance, and momentum (or whatever you choose to call it).
10 Aug 2011
by Danny Tuccitto
Ladies and gentlemen, our Wisdom of Crowds feature returns for its third season. In case you've been living under the same rock as that GEICO guy for the past few years, what we do here is produce crowd-sourced statistical projections for the upcoming season using Football Outsiders' Twitter account to harness the fantasy football aptitude of our readership. This season, we've decided to up the degree of difficulty by focusing on players for whom 2011 stats aren't that easy to predict. Today, let's look at the five quarterbacks we threw out there for Twitter consumption last week.
In years past, we've given reader projections for average fantasy performance, best-case-scenario performance, and worst-case-scenario performance. This year, we'll do the same, except to also add statistical confidence to the mix. Namely, alongside the average projection, we'll give you a range that incorporates the margin of error in that projection. This is basically the fantasy football version of a political poll, after all. Read it as "the average Football Outsiders reader predicts player A to have a stat between X and Y."
Monday: Kevin Kolb
Average: 3,358 ± 110 passing yards , 21 ± 1 TDs, 17 ± 1 INTs
Best-case scenario: 4,164 passing yards, 28 TDs, 10 INTs
Worst-case scenario: 2,500 passing yards, 11 TDs, 28 INTs
Last season, Kolb threw for 1,197 yards, 7 touchdowns, and 7 interceptions in what amounted to 6 full games for the Eagles. If you prorate that out to 16 games, it amounts to 3,192 passing yards, 19 touchdowns, and 19 interceptions. In that context, it appears that the crowd basically thinks Kolb in Arizona equals Kolb in Philadelphia.
Apparently, the crowd also believes Kolb isn't going to perform anywhere close to Kurt Warner's production under Ken Whisenhunt from 2007 to 2009. During that period, Warner -- if you prorate his stats out to 16 games per season -- averaged 4,164 yards, 30 touchdowns, and 19 interceptions. Aside from interceptions, that's on par with what our readers say is Kolb's best-case scenario. With a suspect offensive line and a wide receiver corps that doesn't include Anquan Boldin, that best-case scenario doesn't seem very likely.
Tuesday: Tarvaris Jackson
Average: 2,937 ± 141 passing yards, 17 ± 1 TDs, 18 ± 2 INTs
Best-case scenario: 3,750 passing yards, 26 TDs, 10 INTs
Worst-case scenario: 2,100 passing yards, 10 TDs, 32 INTs
In Jackson's lone season as a full-time starter (2007), his 12-game stats prorate out to 2,548 yards, 12 touchdowns, and 16 interceptions. In Pete Caroll's first season as the Seahawks' head coach, Seattle quarterbacks combined for 3,536 yards, 14 touchdowns, and 20 interceptions. Therefore, readers' average projection seems to be smack dab in the middle of Jackson's history and Caroll's history; that is except for touchdowns.
So the question becomes, "Will Jackson throw for more touchdowns than expected this season?" In our minds, the answer is yes. First off, Jackson's primary targets in 2007 were Robert Ferguson and Bobby Wade, not Sidney Rice and a resurrected Mike Williams. Second, with all due respect to his beast-mode run in last year's playoffs, Marshawn Lynch isn't likely to be the focus of the offense like Adrian Peterson was. Finally, the Seahawks don't figure to be a good team this season, so Jackson will likely find himself in a lot of easy passing situations (read: they're going to be behind frequently).
Wednesday: Donovan McNabb
Average: 3,305 ± 130 passing yards, 20 ± 1 TDs, 14 ± 1 INTs
Best-case scenario: 4,400 passing yards, 27 TDs, 6 INTs
Worst-case scenario: 2,400 passing yards, 11 TDs, 23 INTs
McNabb was a quarterback we featured in last year's Wisdom of Crowds, and he ended up exceeding yardage expectations while underperforming in both touchdowns and interceptions. This season, the prognosticator's dilemma revolves around whether or not McNabb can parlay a better running game into better passing stats despite having a wide receiver corps of equal talent. After all, Percy Harvin may be Santana Moss's doppelganger, but Ryan Torain is no Adrian Peterson.
One clue that might help solve the mystery is that the Vikings replaced offensive coordinator Darrell Bevell with Bill Musgrave. In his last stint in the position, Musgrave coached a Jaguars offense that averaged 3,368 passing yards, 17 touchdowns, and 14 interceptions. With Mike Shanahan last season, McNabb had prorated stats of 4,156 passing yards, 17 touchdowns, and 18 interceptions. Put that together, and you have 3,762 passing yards, 17 touchdowns, and 16 interceptions. That's pretty darn close to what our readers have to say.
Thursday: Andy Dalton
Average: 2,740 ± 131 passing yards, 14 ± 1 TDs, 19 ± 2 INTs
Best-case scenario: 3,500 passing yards, 17 TDs, 9 INTs
Worst-case scenario: 2,100 passing yards, 11 TDs, 28 INTs
By far, Dalton was the least palatable of our quarterback options last week. We received only 21 projections, and more than one reader expressed displeasure with the choice. What? Do you think we're going to make this easy for you?
All kidding aside, reservations about Dalton are understandable. Obviously, he's a rookie quarterback, and that's not exactly ideal. In addition, as we discussed in Four Downs, the Bengals' running game may not provide much relief in Dalton's rookie year. Throw in an incredibly young wide receiver corps, and the readers' average projection looks better and better. One historical note: Dalton's 16-game, average reader projection -- if accurate -- would most resemble the rookie seasons of Rick Mirer (2,833 yards, 12 touchdowns, 17 interceptions) and David Carr (2,592 yards, 9 touchdowns, 15 interceptions).
Friday: Josh Freeman
Average: 3,708 ± 113 passing yards, 25 ± 1 TDs, 13 ± 1 INTs
Best-case scenario: 4,200 passing yards, 33 TDs, 9 INTs
Worst-case scenario: 3,000 passing yards, 19 TDs, 17 INTs
In comparison to the other four quarterbacks of the week, readers have the highest expectation of Freeman. Indeed, he's the only one whose statistical bottom is still a 3,000-yard season and a positive touchdown-to-interception ratio. Not to mention that his average projection is also the best of the bunch.
Compared to his breakout 2010 season, the average reader projects Freeman to have more yards, more interceptions, and an identical number of touchdowns. That touchdown projection seems unlikely. Since the passing game expanded in 1978, only six quarterbacks have thrown for at least 25 touchdowns in two of their first three seasons, and two of them are current or future Hall of Famers (Peyton Manning and Dan Marino). If Freeman manages to join them, we'll be talking soon enough about championships of the nonfantasy variety.
38 comments, Last at 16 Aug 2011, 3:45am by Bobman