Addison, it's perfectly acceptable to take out outliers/anomalies as a separate way to analyze data. If a running back runs 25 times a game for 100 yards, but one of those runs was for 75 yards, that means the other 24 runs netted only 25 yards total (slightly over a yard per play). So yes, an "average" includes all the data, but if you really look at the data (in my scenario) you see that the running back really struggled save one play.
There's nothing wrong with that analysis as long as you don't discount other forms of data. Hell, even professional statistical analysts look at data without outliers or pure averages. True averages can be very misleading.