I’m guessing the data starts in 2000 and sees that season as the first year, so the whole 2000 Pats are included in the data. Just ignore them.

Under the AV system, a player’s AV total will vary significantly for the same performance depending on the quality of the offense or defense that he is on. This is as much a feature of AV as it is a bug when you look at large aggregate groups of players, but when you’re comparing a sampling of player-years from the Belichick-era Patriots with a sampling of player-years from a fairly random selection of the other 31 teams over that period, you’d expect to see the Patriots group have a *much* higher AV if starts, pro-bowls, etc. and all else are relatively equal (to say nothing of the fact that AV provides a multiplier bonus for PB/AP honors and guys on the Patriots are probably more likely to get such honors for a given level of performance due to increased attention and team success).

I’m coming at this as a Patriots fan so I’d love to believe that the players who go to the Pats tend to do brilliantly and the ones who leave tend to do poorly, and there’s a kernel of truth there, but I think taking the AV numbers at face value would lead us to overestimate the magnitude of that effect. You can see that in some relatively dubious head to head AV results: Anthony Pleasant was a solid DE in 2001 (10 AV), but Chandler Jones (9 AV) was almost certainly better in 2016, and I’d say the same of Akiem Hicks (8 AV). Danny Woodhead’s 2010 (10 AV), Brandon LaFell’s 2014 (11 AV), and Brian Waters’ 2011 (13 AV) are probably all inflated a bit due to playing on superb offenses relative to Asante Samuel’s 2008 (12 AV), Aqib Talib’s 2014 (9 AV), and Ty Law’s 10 INT 2005 (10 AV). It’s not so much that those players who left the Pats are necessarily being underrated by AV as it is that some offensive players who were valuable contributors but also somewhat “along for the ride” are getting overrated because there are so many AV points sloshing around on those all time Patriots offenses.