Data Lies: Part I
Aug 2, 2012
More articles from Mark Chamberlin|
Sure, we all want the easy answers. Here’s last year’s data set, here’s how the value broke down by position, here’s where the value was last season, and how you should interpret it for drafting this year. Data is comforting, quantitative and makes analysis much easier. The problem is the data lies, especially in a game like fantasy football. There are simply not enough weeks in a season and too many variables to consider to accumulate a reliable data set. The data is a good starting point, but what you do with that data is what separates your cheat sheet from ‘standard’ to ‘gold.’
First, examine points per game, not total points. This is incorporated into several current rankings, but unfortunately I feel obligated to point it out since it seems over looked too frequently. When you’re starting a player you’re concerned about their weekly output as if they miss a week, you’re not replacing them with a ‘zero;’ you’re substituting him for a replacement level player off waivers or a bench option. Greg Jennings did not finish in the Top 20 in 2011, but he missed the last three games. He finished 11th in points per game, 10th if you omit Kenny Britt and his three-game season. This finish is consistent with 2010 (10th).
Moving on, the above data is easily extracted. The direction we’re going now is where the blanket data set cannot take you. No, I’m not talking about singling out second half of the season stats, road vs. home, or indoor vs. outdoor stats. What I’m talking about takes a lot more time, effort and thought than simply filtering data, which is why many don’t take the time to do it. How did external factors affect players last season? Notably injuries. Adrian Peterson’s pre bye-week stats (21.3 points per game - fourth among running backs) look a whole lot different than post bye-week/pre-ACL tear (9.7 points per game). That’s because he was running through a high ankle sprain for most of it. Personally, I think this practice, while taking a lot of time, is a lot more time efficient than creating projections, a practice many others commit hours and days to this time of year.
The first example that immediately comes to mind, after Peterson anyway, is Tony Romo. By simply filtering the data you see that he was eighth in overall points and seventh in points per game. You also think of his poor record in December (overblown), struggles in the clutch (not relevant in fantasy), and his finish in 2011 is consistent with his draft value in recent seasons. Therefore, he’s slotted in the sixth-to-ninth range of quarterback rankings that is usually drafted sometime in Round 5, give or take a round.
By taking injuries into account when evaluating his 2011, I see that he broke his ribs early in Week 2 against San Francisco. He left in the first half only to find the courage to return late and lead the team to a comeback win. He then played the following week and was clearly affected as he struggled his way to his worst performance of the season against the Washington Redskins (only 10.6 points). Furthermore, in Week 16, he was injured on his second throw of the game and was subsequently removed. Bitter fantasy owners who lost their championship game last year remember this and are partly responsible for his reduced average draft position. This is a value opportunity. If you throw out his two injured performances, his 23.71 points per game quickly becomes 26.34 points, good for sixth overall in points per game and only a couple of points per game behind Matthew Stafford and Cam Newton.
However, you’re getting Tony Romo in the fifth round, sometimes later, whereas a second-round pick is required to get Stafford or Newton. There is more than a two points per game difference between the running back you’re considering in Round 2 than the ones you’re considering in Round 5.
The above example isn’t too difficult to identify and dissect if you’re looking for it. The player was injured and performed poorly, causing his points per game to be deflated - makes sense. Taking the injury analysis a bit further, how did other positional injuries affect a players performance? This portion of the analysis takes significantly more time, thought, and effort; a little subjectivity, too. There are dozens of issues I uncovered, but many of them could be well argued. Therefore, I won’t pick on the not-so-obvious ones. Although some of 2012’s best values may come from those candidates.
To keep it simple, an obvious example comes in the form of Matt Schaub. The popular value pick in 2011 (like Romo this year, fifth round give or take) burned many fantasy owners with his mediocre performances until a foot injury prematurely ended his season in November, spoiling the Houston Texans’ chances for the Super Bowl. His price has been adjusted in 2012 drafts after 2011’s showing to a guy that’s sometimes not picked inside the Top 100. That shouldn’t be the case as he had only three games with his No. 1 passing weapon, Andre Johnson, one of which was against the woeful Indianapolis Colts in Week 1 in which the Texans abandoned the passing game after halftime because they were out to a 27-point lead. His Week 2 and 3 performances in a competitive game and with Andre Johnson? He tallied 23.7 and 34.6 points, the big-time performances that were expected from him.
Then, in Week 4, he lost his stud receiver and he never got to throw to him again as Johnson did not return until after Schaub was down for the count. As you know the weapons behind Johnson on the depth chart (Kevin Walter, Jacoby Jones, Owen Daniels, Joel Dreessen, etc.) were nothing to write home about, so the Texans became a running show, to great success, actually. However, that was not the intended plan coming into the season. With Johnson and Schaub ready for Week 1 again, there is every reason to believe Schaub can churn out the same numbers that caused him to be a popular Top 50 pick in 2011. Only he’s a half dozen rounds cheaper to acquire this year.
This is just two cases (of many) of value picks that the raw data suggests you downgrade that you really should not. Instead of taking the time to accumulate projections that will never come true for your cheat sheet, take the time to dig through the happenings of 2011 and use it to project what to expect for 2012. In the end, your cheat sheet will turn out a whole lot different, and better, than it would otherwise. You’ll uncover value where you didn’t think it was, and your competition is unaware it exists.