Pats averaged 301.8 passing yards per game with him in the lineup, and 226.2 without. Averaged 30.7 PPG with him on the roster and 21.5 without. New England did win without him, but he was hardly a mistake: They were a demonstrably better passing offense with him in the lineup.
This is what I hate about nfl stats. They are meaningless. Think about it, if a team plays like 4 crappy defenses in a row (or is in a division with a bunch of crappy teams) they are talked up as some offensive juggernaut. Every team is different and they don’t play every team each year. There are far too many variables to account for to say one teams performance is because they had one player in their lineup for these games. You can’t break down a team performance without also accounting for the different teams they were playing against, the time of year, injuries to other players, weather, etc. Way too many variables to account for.
I mean, sure, there's always the chance for extreme outlying cases to come into play. That's why it's worth getting a decent sample size, and 10 games vs. 9 games is, generally speaking, a pretty reasonable selection size. Your odds of the ten games that one player stuck around for just being really shitty teams, and the 9 games he didn't being defensive juggernauts is pretty small, on average. The point of the post is more to demonstrate that there's no grounds for calling him a mistake: Even in a limited sample size, there's no evidence that he made New England worse.
That being said, to add some context, here's the passing defense ranks of the teams NE played with and without Gordon.
With: 21, 16, 31, 7, 1, 12, 6, 24, 3, 21
Without: 28, 2, 8, 10, 24, 9, 31, 14
Not a ton of variation there. Obviously things can change from one part of the season to the next, but again, the point is more that there's no reason to believe Gordon was a detriment.