aghcsm
Well-Known Member
I've been interested in the fairly routine point differential shown by the Buffs in home-and-away (HAD) conference games, and wonder if it might be a useful tool for projecting likely results and evaluating performance. Specifically, the average differential (starting with PAC12 play) has been 19 pts in 2011-12, 14 pts in 2012-13 and 18 pts in 2013-14.
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.
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.
