The example of being able to run for a couple of 1st downs at the end of the Viking game is an outlier - outliers are just that, and have to be dismissed. They are called outliers b/c they lie outside the set of observed data that give you your trend and are therefore unreliable and invalid.
wist