According to the source below, the partial conversion rate is about 57 gallons per acre (at 39 bushels/acre). It takes WAY less than 57 gallons of biodiesel to farm one acre of soy. Modern practices are really only going to have 3-4 passes (no till plant, early spray, maybe a later spray and harvest). After that is is mostly bulk trucking which shouldn't be much per acre.
Remember that unlike corn, soy requires very little/no fertilizer due to it being able to fix its own nitrogen. It is mostly the fertilizer the kills corn.
Cane ethanol seems to do better than soy fuel. I'm also aware that new biofuel "generations" have the potentional for much higher returns, but these things are either still emerging technologies (e.g., cellulosic and algae) or horrifically damaging to the environment (e.g., palm oil biofuel).
I was with you until this. a brief thought experiement makes it seem impossible that it would be 1 to 1, but rather that some smaller amount of fossil fuel used would yield a much larger amount of biofuel
The value of biofuels is simply cushion in case of poor harvests.
There's a lot of debate about this, with numbers varying due largely to differing assessments of inputs and credits for byproducts, but I have never seen a convincing EROEI for corn ethanol over about 2.
EDIT: I forgot to include the link! Here it is: http://netenergy.theoildrum.com/node/6760