Comments on Advanced NFL Stats: Fantasy Draft Strategy

Good stuff. Basically what you spoke about correl...

2008-08-18T22:49:00.000-04:00

Good stuff. Basically what you spoke about correlates to AVT Theory which is what good managers rely on early in drafts. The key is projecting the right guys to the numbers. That is where the difficulty lie. In the end, known commodities should be taken in the very early stages with upside guys later

Sky-On your second point, you're right. That's exa...

2008-08-18T16:16:00.000-04:00

Sky-On your second point, you're right. That's exactly what regressing does. It accounts for the consistency levels.

You are also on the money about kickers. I'm not trumpeting that I've discovered something entirely new, just quantifying what most people intuitively understand already. People seem to get VORP when drafting kickers, but look at absolute scoring when picking TEs.

Omar-by player. Basically, I want to how consisten...

2008-08-18T11:28:00.000-04:00

Omar-by player. Basically, I want to how consistently do elite players repeat the next year.

Sky-No. I think replacement-level is a good and us...

2008-08-18T11:09:00.000-04:00

Sky-No. I think replacement-level is a good and useful concept. What I'll eventually get to is a specific version of replacement level. What is *your* replacement level at *this point* in a draft. Because everyone in the end is forced to have the same position-based lineup, the focus should be on maximizing the drop-off between your picks and everyone else's.

Did you look at the correlations by player or by ...

2008-08-18T10:59:00.000-04:00

Did you look at the correlations by player or by ex post rankings?

Since regression to the mean is the norm for NFL t...

2008-08-17T16:59:00.000-04:00

Since regression to the mean is the norm for NFL teams (ie: overall record), then wouldn't drafting players on teams that should regress from below .500 while avoiding players on teams teams due to regress from above .500 last year be a good strategy? There seems to be so much randomness in individual performance, given random injuries and such.

Are you eschewing replacment-level theory here? I...

2008-08-17T16:19:00.000-04:00

Are you eschewing replacment-level theory here? It's by far the most important way to take account of scarcity. Players should only get credit for points they'll produce above the likely last picked player at their position. That's why the best kicker is worthless even though he scores 150ish points.

Also, is consistency really that important if you've properly regressed a player's projected point total?