When I put together this week's series of articles on pending free agents written from the perspective of a free agent "book," the intent was to make the argument for a player in a way that highlighted his strengths while downplaying or even ignoring his weaknesses. I didn't make up any of the facts or neglect to include any players who might make the comparisons I made look worse, so it would be unfair to say that the series was dishonest, but it definitely distorted statistics and context in ways that made the players in question look more valuable than they actually are. And that's not an apology, since there was a disclaimer on every article. Believe it or not, agents actually use statistics in similarly evil ways. Shocking, I know.
So now that we've all enjoyed a week of inflating player values, it seems like a good time to quickly highlight the weaknesses in those analyses and identify the arguments that negotiating teams might use to try to drive down each of the four free agents' value. The existence of these counterarguments doesn't make the initial packets that we put together irrelevant, but it should identify some of the weaker logic points. As with the first set of articles, though, these takedowns are strictly based on the perspective of an organization, not my own. My feelings about all these players, truthfully, lie somewhere between the agent-friendly vibes and the organization-friendly ones you're about to read.
- The Steelers' slow pace came as a result of their passing game. Because Pittsburgh had the fewest possessions in football on offense, I suggested that Wallace would see his numbers increase with more possessions on a faster-paced team. That's probably not the case. The Steelers almost surely had such a low number of possessions because they threw the ball more frequently, which led to longer drives and fewer three-and-outs. Alongside Pittsburgh with low drive totals were San Diego (166), Green Bay (168), and New England (173), all of whom were well below league average in possessions but threw the ball (and scored) at league-leading paces.
- He's gotten a competitive advantage by playing alongside Ben Roethlisberger, not been disadvantaged by Roethlisberger's absence. Sure, Wallace's numbers are better with Ben Roethlisberger in the lineup, but everyone's performance is better when their starting quarterback is around, especially if he's as good as Roethlisberger. Plus, what happens if he ends up going to somewhere like Jacksonville or Kansas City and has to play with a below-average quarterback all the time? The important variable is "Roethlisberger," not "starting quarterback."
- He's way worse than Randy Moss and Torry Holt. I noted that Wallace was probably the best prospect since Moss and Holt, and while that might be true, both Moss and Holt put up far superior numbers to Wallace's totals during their first three seasons in the league:
- The definition of a "great game" by a quarterback is totally arbitrary. Why does a great game from a quarterback need 400 yards? Because it fits that one Matt Flynn start, I suppose, but has a quarterback who has thrown for 380 yards instead of 405 really played a far worse game? Of course not. If we remove or change even one of the criteria, our list of all Pro Bowlers begins to fade away. For example, if we simply ignore the criteria that the team needed to win the game, it adds Tommy Maddox and Tony Banks to the list. That's a far less appetizing group.
- Flynn's big game against the Lions was in Week 17. It's reasonable to suggest that the Lions weren't incredibly motivated to win that one.
- The no. 1 pick he played behind at LSU was JaMarcus Russell. Considering that the LSU coaches managed to convince the world that Russell was worth the first overall pick, well, maybe they're the ones who deserve the hype article.
- The vast majority of players with Flynn's lack of experience through 26 failed to turn into viable pro quarterbacks. I mentioned a few exceptions in the piece, but of the 40 or so quarterbacks with a similar professional record to Flynn, only those exceptions managed to get past 600 or so additional pass attempts.
- Flynn doesn't get to bring the Green Bay receivers along with him. He didn't have Greg Jennings in Week 17, but he did have Jordy Nelson and James Jones. That's better than what most teams have to offer as their starters, let alone the no. 2 and no. 3 guys.
- Williams has injury issues. In addition to his torn pectoral muscle from 2011 — an injury that recurs easily — Williams suffered a hernia in 2010 that cost him the final few games of the year. Including the two playoff games from this year, Williams has only played in five of Houston's last 16 games.
- Houston's defense wasn't very good with him and was very good without him. It's often too easy to note a change in the availability of one player and link it directly to a change in team performance, but Williams was unable to push Houston's defense into even the middle of the pack during his first five years with the team. In year six, Wade Phillips's defense was excellent with Williams in the lineup and equally good without him, as the combination of Connor Barwin and Brooks Reed combined for 17.5 sacks at outside linebacker.
- All the cumulative stat tables were designed to account for Williams's absence due to injury, but nobody else's. For the years where Williams stayed healthy, I used cumulative tables that highlighted his raw sack totals without accounting for his games played. When I started to include the years where Williams was struggling to stay healthy, I moved to sacks-per-game statistics that ignored his mediocre rookie season (4.5 sacks in 16 games) while making his partial season in 2011 (five sacks in five games) look like a dominant one.
- Foster's playing in a zone blocking scheme that has a history of producing successful seasons from unlikely candidates. Remember that the likes of Mike Anderson and Olandis Gary went from afterthoughts to 1,000-yard backs overnight in Denver under this same blocking scheme. Foster plays in a great scheme with dominant offensive linemen in front of him. That's incredibly important in discussing what he can and can't do.
- He's beginning to develop a fumble problem. It could just be randomness, but Foster fumbled six times on 331 touches during the regular season and then once more during the playoffs. There were also a number of close calls where it looked like he was struggling to hold on to the football, even if he did successfully manage to do so.
- Not all yards are created equal. While Foster can be a useful receiver out of the backfield, receiving yards for running backs depend heavily on context. Is a team splitting out a back and actually having him run patterns against overmatched linebackers, or is he just a dump-off option when nobody's open downfield? In Foster's case, especially with T.J. Yates under center, it was more the latter than the former.
So now that I've pointed out how the stats can be exploited and manipulated, does this mean that you can never trust football statistics (or me) again? I'd hope that's not the case. The examples above are ways to use statistics incorrectly, but there are still ways to use them in a correct, comprehensive manner. As a general rule, the more arbitrary a stat seems, the less meaning it's likely to have. And while it's easy to come up with one anecdotal stat to point out a player's strengths or weaknesses, it's important to consider the bigger picture. Has this player exhibited that level of play for multiple weeks or even seasons? Does the stat in question consider every aspect of his play? Does it consider the game situation he's in when he's accruing these statistics? Does the comparison being made between one player's statistics with another make sense on a level beyond mere stats? Did he play in a system that would make similar players to the guy in question look good? A good reader — a good football fan, really — should be skeptical of statistical analysis without being dismissive. And a good analyst provides the context needed to unlock the right answers to those questions.