So, if you throw out a long play as an outlier, how many of the plays where he lost yards do you throw out? Don't forget, he was hit in the backfield on the play that he went for 80s yards and that could easily have been a loss. In statistics, you don't just throw out one data point--you throw out a number from both sides. And you do it for every data line--not just one, or one player in this case. It isn't right to take away Brown's best run and then not do the same for every other RB if you want to compare their ypc.
The other thing Rotoworld didn't consider: offensive scheme. In 2011, once Manning went down, they finally started running more I formation. So, including Brown's previous ypc and saying "that is more indicative of his ability," is misleading. I have argued before that the shotgun draw plays that IND ran almost exclusively with Manning was not beneficial to Brown's running style. A big reason he did better in 2011 was Manning going down and a more run friendly offense being used. To be fair, we don't know how Brown will run in yet another new offense in 2012, but I am pretty sure it will be better than the shotgun draw of the past. I don't expect him to gain 4.8 ypc this year, especially if his carries goes up, but I wouldn't be at all surprised to see 4.5 ypc. Another variable? The offensive line. In Brown's early years the IND line was built to pass block, not run block. And they didn't open many good holes. The OL was revamped this year with bigger OL. I think that could bode well for Brown as well.
Bottom line: if you want to throw data out you need to do it more consistently (all players) and systematically. This web site offers a basic approach:
http://www.ehow.com/how_5201412_calculate-outliers.html