Statistical Anomalies and the Occasional Misplaced Victory

There are certain players who can, from time to time, place fantasy teams on their backs and give owners a misplaced victory.  I'm not talking about the LaDainian Tomlinsons of the world.  Everyone knows to target players who turn in great performances or super-great performances or super-double-plus-great performances.   I’m referring to the second- and third-echelon players who go in the later rounds because, even though they do produce breakout games, they are discounted as “inconsistent” since those breakout games are so few and far between.  As crazy as it may sound, we often end up holding breakout games against players just because we’re disappointed that the breakout games don’t happen more frequently.  If you went into the season ranking Donald Driver ahead of Plaxico Burress (as I did), you’re guilty of just this sort of wrongheaded thinking.   

I remember back in 2002 when Plaxico Burress went supernova against Atlanta with nine catches for 253 yards and two deep scores.  According to a fairly common fantasy scoring system, that performance racked up 52 fantasy points.  However, in Week 2 of that same season, his dismal performance against Oakland netted him -1 points according to the same scoring system.  Those games led many players to conclude that Burress was too inconsistent to be relied upon, but that overgeneralization doesn’t hold up to scrutiny whether we look at Burress’ breakout games or his ordinary games or both, as I hope to show with hard numbers in what follows.   

This past weekend, Burress destroyed the Cowboy defense—and gave many fantasy owners victories that they might not have deserved.  My first thought was, "If that was the upside, will there be a commensurate downside forthcoming?" 

I decided to do some checking against three other very "similar" WRs—similar with respect to experience and total fantasy production over the past five years.  I tallied all the points scored since the beginning of the 2002 season, inclusive of Week 1 2007, for Burress, Donald Driver, Santana Moss and Laveranues Coles (that's 81 scheduled games each).  The results surprised me, especially when considering the upside vs. downside of each player. 

Burress didn't play in seven games, so we have a sample of 74 contests.  He's tallied 982 points, giving him an average of 13.27 per game.  He turned in a nauseating 11 games of four points or less (don't forget one of those was actually less than zero), with three of 40 or more (43, 43 and 52).   

Driver missed only one game in that stretch, offering a grand total of 1029 points for an average of 12.9.  He matched Burress' dyspeptic performances (11), but has never scored over 40 or more (in fact, he's only scored in the 30s three times). 

Moss gives us a sample of 76 games, a total of 955 points, averaging 12.57 per showing.  He's turned the stomachs of owners a disgusting 19 times (exactly one fourth), but came up with monster performances five times (40, 41, 41, 46 and 49). 

Coles gives us 79 starts with 945 points, averaging 11.81.  He's barely broken 30 two times, never even scaring the 40 point barrier, while he's logged 13 games of four points or less. 

I picked these four players because they have fairly similar numbers over careers that couldn't be called inadequate.  In fact, during this season's draft these players were typically taken within half-a-round of each other.  There isn't much that really distinguishes them.  Even their point totals (982, 1029, 955 and 945) over five years are incredibly close.   

These players achieved these numbers in very different ways.  And it is important for fantasy owners to temper their expectations by understanding the "hows."  While it is tough to argue with Driver's and Coles' consistency, I promise that Moss and Burress have single-handedly won more games for their owners.  And while Moss' downside is disturbing, Burress' isn't.   

Going one step further, let's eliminate from our samples all breakout games.  Moss is left with 779 points over 72 games, giving a 10.82 average.  Burress brings 844 points over 71 games, giving an 11.9 average.  So if we compare Burress’ average exclusive of his breakout games (11.9) to Driver’s average (12.8), we see less than 1 point of difference between them.  When you consider that on at least 3 occasions, Burress won games for his owners almost single-handedly, it’s difficult to argue that the touted “consistency” of a player such as Driver is all that important.   

In this light, Burress' downside is really a non-factor.  His average is still very close to that of Coles and Driver, but with him comes the occasional undeserved gift. 

I've had plenty of "consistent" teams in the past.  Those teams have consistently lost in the playoffs, as well.  While minimizing the downside is certainly critical, owners should always look first for the upside.