So in every test that I have done, in isolation the reactors perform exactly as they should. Ignoring the debug values and going by adding up core transfer numbers, any combinations of reactors work just fine. I have to conclude that (again) there's a critical bug in stock to work around. Here's some testing of my own. This base has enough radiators deployed to run the reactor at full power (800 kW) plus 4 drills at regular power or 2 drills at max power (200kW) 1) The reactor is