It does incorporate SoS, but if you look at the data sets, you will see that it's not devalued enough. SDSU ranks super high (top 20) on DFEI but has the #122 SoS. If you look throughout the rankings, you will see that average teams with low SoS strength routinely fare better than average teams with high SoS strength.
For instance, of the current top 25 in DFEI, 11 have SoS ranked worse than 40. Two, SDSU and Troy, are 126 & 128, respectively.
Also, SEC teams get a bump in SoS, even though fellow SEC offenses are generally poor. My argument is that SoS rankings are themselves subjective and volatile, and assigning an arbitrary value to them for the purpose of equating DFEI is problematic.
So, again, all this to say that while I do believe DFEI is a valuable stat, I also think it has inherent flaws. Most statistics do. Only when you compile several stats and compare them can you truly get a better idea of where a team is.
@AE 87 basically said the same thing in his post above.