He should've omitted the shortest play as well. When you take away the largest and the smallest gain and average the rest, you get an accurate gauge of the player's performance during the game. It's a standard practice in statistics.
which does not translate into football. When a runningback gets a bunch of carries, he will likely have multiple short gains of less than 3 yards. Thats part of grinding it out on the ground. You pound away with short gains until eventually the defense gets worn down and the RB busts a long one. But 2 or 3 yard gains are much more common than a 40 yard gain, so removing the longest gain and the shortest gain is going to pull the ypc total downward in almost every case.