how exactly do you incorporate the defenses of the teams into this equation?

Obviously the predicted offensive output is equal to the opposing team’s predicted defensive effort.

In a given area, say passing, the offense and opposing defense are each defined as being under, or overachievers. If, for example, a team averages 100 yds passing against teams that usually give up 150 yds, the team is an underachiever.

Achievement is quantified as a ratio. On offense a ratio 1 overachievement. The reverse is true on defense. I average these two ratios.

Then I average the team’s typical pass yards with what the opponent typically gives up on defense.

Then I multiply the two averages. That is the predicted value.

Or consider this….
T1PO => team 1 avg pass offense
T1oppPD => team 1 opponents avg pass defense
T2PD => team 2 avg pass defense
T2oppPO => team 2 opponents avg pass offense
T1PVPO => team 1 predicted value pass offense

T1PVPO = .25 * (T1PO/T1oppPD + T2PD/T2oppPO) * (T1PO + T2PD)

wow and here I thought grad school was time consuming and hard.

I only say that so people don’t expect too much from me.

What sigma level are your predictions falling into? How about other factors such as home field….do you have any adjustment factors for this? Have you done any log graphs…I only ask because I like logs…I used to live in one.