I'm not ignoring the 3 out of 8 cells returned as depleted isotopes from the centrifuge... I never factor in 8 cells per centrifuge run to begin with.
My math assumes that you insert 5 re-enriched isotopes into the centrifuge and get out 1 plutonium and 4 thorium. True, in practical application you have to insert 8 isotopes and get a side product of 3 depleted ones, but those can be re-bred indefinitely. Therefore you don't actually lose or gain any reactor fuel, and the whole thing can be safely ignored. The only thing that counts is that through some process, the details of which are unimportant, the centrifuge converts exactly 5 isotopes into exactly 5 fuel cells with no loss beyond the cost of 100k EU per centrifuge run.
I'm also not factoring in breeding fuel costs because really, silktouching a single coal vein and throwing the blocks into an indsutrial grinder will give you enough thorium to breed hundreds upon hundreds of isotopes without ever touching your regular reactor fuel at all. Breeding is essentially free of charge beyond the initial build cost of the breeder reactor.
And yeah, that centrifuge recipe is very useful. Together with the fact that the industrial grinder will double uranium output from ore plus guaranteed plutonium dust, and gives guaranteed thorium dust from coal ore (you version's mileage may vary), you can still get a decent amount. You do however invariably end up with uranium cells that you need to run through a regular reactor.