Averages are just statistics of course, and individual games may be affected by injuries, suspensions, 'off-nights' or 'on-nights,' special motivation etc. There was one HAD each in the 11-12 and 12-13 seasons where the Buffs did better in the road game. However, those rare exceptions serve to support the concept. The way I envision using HAD is as follows: example 1 - Buffs beat USC by 21 at home, if the typical differential is taken as 18, they would expect to win by 3 away (actual win was 4 and a 17-pt differential. Example 2 - Buffs beat UCLA by 6 at home, indicating an expected 12-pt loss on the road (actual loss by 13). Examples 3-5 - Buffs lose by 23 at Utah, estimate loss by 5 at home; lose by 14 at Arizona, estimate win by 4 at home; lose by 6 at ASU, estate win by 12 at home. Again, there are many factors and the numbers no doubt often will be way off - but IMO it's a nice starting point for expectations and evaluating whether the team performed as expected, or better / worse than expected. When the numbers don't match it's a good starting point for considering 'why.' I certainly wouldn't expect the Buffs to lose by 20 at UDub, due to multiple issues. Interested in thoughts on this.