The problem with your reasoning is it doesn't account for 3rd, 4th, 5th order effects. One system tweaked in such a way effects another sytem in a way that would need to be compensated for. The precision required would have made formal documentation absolutely necessary. Tweaking a system on race day and it working better than expected and not being able to reproduce the result is far more possible than doing the same thing only orders of magnitudes more complex. My favorite are the original Apollo tapes that have impossibly fast communication between astronauts and ground control.