As you've shown, deck 2 is the best of the lot. But this was not your thesis. Your thesis was that Deck 2 is more of an improvement over deck 1 than deck 4 is over 3.
That was not my thesis, but I could prove that as well, if you'd like. First, however, a few points:
1. You assumed that there were an equal number of Coppers and Estates/Great Halls in each deck. I made no such claim. Furthermore, that doesn't matter, given that I was talking about maximum output, not average output.
2. With your 2.5 vs. 5 arguments, you seem to assume that Dominion's cost scale is linear. It is not. 5 is more than twice as good as 2.5
However, let's go with your assumption that we have a deck that starts with an equal number of each type of card (at least 5 copies of each).
A deck with only Coppers and Estates averages $2.5 per turn, and can range from $0 to $5.
A deck with only Coppers and Great Halls will always generate $5 per turn.
A deck with equal numbers of Coppers, Estates, and Scouts averages less than $2.5 each turn, and can range from $0 to $5.
A deck with equal numbers of Coppers, Great Halls, and Scouts is likely to draw the entire deck on some turns, generating as many coins as there are Coppers in the deck.
Hopefully we can agree that $∞ is a greater improvement over < $2.5 than $5 is over $2.5.
EDIT: I retract my claim that the Scout/Great Hall deck is in any way likely to draw itself entirely. That's not at all likely unless about half of your deck is Great Halls.
That being said, you're still fairly likely to reach $8.