What you mentioned is correct as far as the ECU using a bunch of voltage signals from a whole host of sensors to make decisions. These signals should be as interference free as possible (aka have the best signal-to-noise ratio possible). To achieve this, the engine is thoroughly grounded from the factory. Adding more grounds to something that is already grounded will accomplish absolutely nothing. The stray electrons will take the path of least resistance to ground. If you add more pathways to ground (think grounding kit), one of them may (just by chance) be a quicker path so the electrons will choose that one. What will that do? The electrons will make it to ground 1 trillionth of a second faster, aka nothing.
In addition, there is a certain amount of "uncertainty" with the sensor signals so any minor interference from this sort of thing will likely make little or no difference to the RMS (quadratic mean) signal that the ECU uses. The increase in signal-to-noise ratio using a grounding kit would be very small.
The only instance where additional grounding may be useful is when you modify the grounding by, say, installing a non-conductive gasket that creates a metal "island" that contains an important sensor where electricity can not "escape" to ground. In this case, it is possible that an electric field could build up modifying the sensors potential thereby sending an incorrect signal to the ECU.
What's the likelihood of this? Basically nill. What would happen if it did? Basically nill. The change in sensor signal would likely be within the normal "uncertainty" and ECU operation would be unchanged since the change in the sensor's RMS signal would be virtually nothing.
That's my take on it. I am no expert but I have a BS in physics and two years of graduate work in mechanical engineering under my belt that should count for something! :biggrin: