The fact that different Demigods have different strengths and weaknesses?
End of Epiphenomenon's quote
Yes, we know that. How does that help prove your point? Lets say DG build X1 is strong in aspect A and weak in aspect B. DG build X2 is strong in A and weak in C. DG Y1 is strong in B and weak in C. DG build Y2 is strong in D and weak in E. DG build Z1 is strong in C and weak in B. How does that information tell you anything about which DG counters which?
Answer: it doesn't, and in fact does not even support the conclusion that such DG X < DG Y relationships exist in the first place.
Take the actual example I gave. I really can't see how anyone other that Oak could counter 4 Reguluses. Four Oaks (it could be pulled off with 3 or even 2, probably) actually do a really good job countering 4 Reguluses. If my opponents switch to all Oaks, there's no way in hell I'm going to stay with 4 Reguluses. It's no longer an optimal combination. I'm going to switch to something else. There is too much homogeneity with four of the same type, even with different builds. This allows a specific counter that will be better, because there are certain things that each specific demigod *cannot* do, even with differentiated builds.
End of quote
Lets talk about 3v3 since thats the standard format and baseline. Lets say the enemy team has 3 Regs. Number of required Oaks = 0. Number of Required *Any Specific DG* = 0. So lets say your team jumps into my game. I have about five 3-DG combinations that I am especially fond of, none of which involve doubles. Your team picks 3 Regs. My team does not change anything and proceeds to stomp you anyways because 3 Regs is bad.
Additionally, team automatch features generally do not let you see your opponents pick of race/character/etc before you are actually playing. I assume that this is how it would be done for a DG team automatch, so this behavior would never even be possible, let alone probable.