Week 3 fantasy matchups

It's time to consider Cam Newton a viable fantasy starter at quarterback. Mark J. Rebilas/US Presswire

My first foray into formally writing about fantasy football matchups ended up being a mixed bag. Players like Joe Flacco, Tim Hightower, Jahvid Best, Jared Cook, Carolina's Steve Smith and the Kansas City running backs held up their ends of the bargain, whereas others like Cedric Benson, Chad Henne and Johnny Knox didn't -- to put it mildly. Just as NFL teams scout their weaknesses each week, it's useful to consider what went wrong with some of the picks.

The most obvious culprit is the statistical uncertainty that surrounds offensive and defensive performance this early in the season. Such uncertainty is exactly why I told you to basically ignore matchups in Week 1. As the season goes on, we gather more and more data, sample sizes increase, estimates of team strength become more reliable and fantasy matchup analysis gets more reliable.

Take Henne's flop against the Texans. There were all kinds of contradictory statistical indicators involved in that matchup. According to Football Outsiders' DVOA projections going into the season, which tend to be pretty accurate, Houston wasn't supposed to have a good pass defense, regardless of whatever magic show Wade Phillips and Johnathan Joseph brought to town. In Week 1, they dominated a Colts offense led by Kerry Collins, which was about as difficult as evading a sheriff's office led by Roscoe P. Coltrane. On the other end of the equation was Henne, who we projected as being better than most people expect, and who had just come off of a 400-yard passing game.

Now, throw into the mix that our matchup analysis gives a bonus for players at home, and that the Texans-Dolphins game was in Miami. And before you say, "The Dolphins were 1-11 in their past 12 home games," remember that (A) that's Duke-capture rare for an otherwise average team; (B) Henne had just thrown for 400 yards at home; and (C) even if the Dolphins are truly a bad team at home, that could mean they're more likely to pass while losing.

So, putting it all together, what you had in Henne's case was a profile that suggested a good matchup, but was nevertheless awash in statistical uncertainty. In this particular instance, the statistical analysis ended up being wrong. However, if that game took place in Week 12, we'd be much more confident about the true nature of Houston, Henne and his home-field advantage in 2011; and therefore wouldn't recommend him.

Therefore, just to reiterate, although I hit on several good matchups -- including the Smith one, for which I caught some grief in last week's comments section -- these analyses get more and more accurate as the season progresses. At this point, my advice is to use them for decisions at the margins of your lineup. If you're between two players, and one shows up here as having a particularly bad matchup or a particularly good matchup according to the statistics, then pull the trigger. Otherwise, keep going with your main guys until further notice.

With all of that said, you guys expressed some confusion about how to interpret the ± point totals. First, I'm working off of a scoring system that gives 1 point per 20 passing yards, 1 point per 10 rushing or receiving yards, 4 points for a passing touchdown, 6 points for a rushing or receiving touchdown, minus-2 points for an interception and minus-2 points for a lost fumble. Second, the actual ± value represents the number of points a player is likely to gain or lose because of his matchup when compared to an average player at his position facing an average matchup. For instance, Wes Welker's plus-5 this week means his matchup against Buffalo suggests he'll score five more points than an average wide receiver facing an average matchup (think Brian Hartline at Cleveland).

Now that we have all that out of the way, on to the best and worst fantasy matchups of Week 3.