Finally done with first pass of projections

So I have finally completed the first pass of player projections (only about 4 months later than I had planned). All 32 teams can be found on the right-hand column under "2008 Team-by-Team Projections."

The next step is to take these raw projections, and adjust them for schedule. For those of you that haven't been following the site for very long and don't know about our True Player Value system, here it is again in a nutshell:

Each team's offense and defense are given a rating based upon how they fare against the league average. Since nobody plays a league average defense every week, the offensive stats have to be adjusted for the defense they are playing against (i.e. an average performance against a good defense is actually a good performance). And the defensive stats also have to be adjusted for the offense they are facing (i.e. a good performance against a poor offense may just be an average performance). If you think about it too hard, there seems to be a bit of a chicken-and-the-egg thing going here, but we proved in the original article that there is a complicated matrix equation that can be solved.

So we solved the equation for last season's games. A value of 1000 is average, with larger values being bad (1100 means the team gave up 10% more than average in that category), and smaller values being good (900 means the gave up 10% less than average in that category).

When projecting defensive ratings for next season, I took these final 2007 values, looked at what the team added and lost defensively, and came up with adjusted 2008 projections for yardage given up. For TDs given up, I adjusted the values to something "normal" for the yards given up value. It turns out that teams generally attack an opponents weakness in the red zone, so a team with a poor rushing yardage defense like Oakland (1354) will have an even worse rushing TD defense (2166). Since teams attacked them near the goal line so aggressively on the ground, the Raider's better than average defense against passing yardage (903) becomes even better against passing TDs (739).

All the defensive ratings I came up with are shown in the 2008 projections for each team. So now its just a matter of going through each team's schedule and adjusting the weekly stats for each player, right? Well, not so fast.

Since most leagues (and every league I play in) have their championship week in week 17, I won't be including that week's stats in my projections. Why draft a running back based on projected stats if the 150 yards he is slated for in week 17 are useless?

So we get rid of week 17 stats, and suddenly I am projecting about 6% lower total stats than any other site. That might not look right to the casual observer, but more importantly, I think that there is a place where we can make up that 6%.

Let's face it, the goal of playing this game is to make it to the championship game and win it all. So, let's assume that anyone reading this site and using these projections will make the fantasy playoffs. I have used this system myself for years and nearly always make it. But how many teams rode the Dallas Cowboys offensive players to the playoffs last season only to see their team get crushed in the playoffs when the Cowboys were shut down by the tough Philadelphia Eagles defense? My money is on a lot. So we need a way to bias players that have favorable matchups in the championship weeks, which are weeks 15 and 16 in most leagues. We do this by counting stats accumulated in weeks 15 and 16 50% more than stats in every other week. Thus, we have a draft list that is biased towards players with good matchups in championship weeks, does not include stats for weeks that are not used by our leagues, and still accounts for a total of 16 games worth of stats.

So, hopefully I can get this done quickly over the weekend and have team ratings updated by Sunday night. I'll let you know when I'm done.