I'm getting a Mean Absolute Error of 11.10 points on my NFL totals model (on a fairly sizable testing set) comparing against the actual scoreline. It seems to me that the market's MAE over the last 4 years is 10.4 points on NFL totals.
I've not gone to the point of testing ATS, but given my MAE result on my out of sample testing set, can I expect to fare well on totals ATS?
I've not gone to the point of testing ATS, but given my MAE result on my out of sample testing set, can I expect to fare well on totals ATS?