I don't want to think of you as anything but a quality poster. That's all I want to think of everyone, although, sadly, it doesn't play out that way. However, when people talk about taking games away when an entire season was played, it's pretty much always an attempt to warp the story while pretending to do otherwise. This is particularly true when you're talking about a team that was faced twice during that season.
You didn't talk about ditching the Giants' numbers, for example, which were 19 carries for 46 yards, or even the 8 carries for 18 yards against the Steelers. If you ditch those games, the numbers go up. That's just how statistics work.
Every game skews the numbers, whether it be to skew them higher, lower, or more firmly into the middle. It's just how data input works.
138 for 584 is what you get when you take away all 4 of those games, by the way, which would be a 4.2 average. It's not as if his numbers drop off a cliff when you remove the 'anomalies'.
P.S. If you want to remove weak teams, get rid of Denver for BJGE. The Broncos were the second worse run defense in the NFL, ahead of only Detriot. If you take Denver out of the BJGE data, you get 61 carries for 210 yards and a mere 3.4 ypa.
I mean, let's be fair.