There's more to statistics than dividing the number of incidents by the number of dives. In particular there's ways to compensate for unknowns or fragmentary data and to get a value for how much confidence you should put in the result.
I have not seen Andrew Fock's presentation (I was comatose in my hotel room :brickwall), and I haven't read his paper (is there one?), but until I read a
proper statistical argument to the contrary, I'll go with that. No offense.
Back to mCCR vs eCCR, I don't have relevant experience with either but I have been in software for a while, including some (low level) aircraft stuff.
When it comes to arithmetics, keeping a watchful eye, or fine control, I'll take the computer every day of the week. A computer does not get bored or tired. It does not forget. It doesn't feel pity, or remorse, or fear.
Sorry, got carried away here
Sure, sometimes there's problems (Therac 25, Ariane 5, Mars Climate Orbiter...), but on balance, it's much safer.
It's not me saying it. It's medical device manufacturers. It's Boeing and Airbus. It's the engine manufacturers, most likely including the one in your car. Along with your ABS brakes.
And for when it does go wrong, we're supposed to have practiced that, and that applies equally well to both.
Cheers,
Matthieu