It is a logical deduction from physics: static friction is generally higher than dynamic friction - you don't want the wheels to spin and end up in the dynamic area.
It may well be a logical deduction on the surface, and at first glance it looks reasonable, unassailable even. However, it's obvious to me that a rally driver can get a car around a loose surfaced rally track, using all those very messy looking rally driving techniques that so very clearly involve dynamic friction, much faster than someone could get around the same track using tarmac driving techniques, and tarmac driving techniques adhere (no pun intended) to your tenet of avoiding dynamic* friction at all costs.
So basically there is more to it than a simple first order analysis will discover. What that is exactly, I don't know, but the methods adopted by rally drivers would strongly suggest that the art of maintaining control (with the greatest overall speed) on loose surfaces come down to more than simply maximising 'grip'.
Full disclosure: I have some little tarmac racing experience, a long time ago and was an RAC/ACU racing license holder (That is not as impressive as it sounds, it basically means you are permitted to race. They give it to you so that they can take it away again if you're a bad boy.). On the other hand I haven't a clue how rally drivers do what they do, despite having been navigator for a mate on some rallying** (it basically involves knowing where you are and reading pace notes off to the driver, you don't need to have a clue what's going on beyond that). All my driving and racing experience has been of the type that drums into you keeping the forces on your tyres within the envelope of their [rolling] static friction. I suspect that even if someone now showed me how to drive like a rally driver I wouldn't be able to make myself do it, backing off the second I hit the edge of that envelope.
*I'm sticking with SilverSolder's terminology even though it isn't quite kosher, and because without looking it up I can't remember the kosher terms. Tyres are rolling, so 'static friction' isn't quite the right phrase, neither is 'dynamic friction', but I think we all know what he means. I do remember that tyre dynamics is not a simple subject, and that tyres only 'roll' in a straight line and that dynamic forces involve 'deformation', 'slip' (which doesn't mean slipping) and lots of other terms that I have forgotten.
Edit:
**Just to drive home how unimpressive holding an RAC/ACU licence was, I seem to remember that I got roped in precisely because I did have a licence and navigators were required to have one for the class(es) he competed in. My license was obtained to lose 2 stroke motorbike races on tarmac, but it wasn't a licence for that in particular, it was just a license to participate in racing. So if anyone ever tries to impress you with having held an "RAC/ACU racing license", you have grounds to be very sceptical that it implies any particular proficiency.
I agree with everything you said about tarmac driving and about friction.
However, my opinion is that there are further nuances to rally driving.
I do not hold any tarmac racing qualifications, although I do hold a regional rally permit. The permits here carry about the same weight that you describe; basically I should know what I am doing and am expected to act & drive responsibly.
The fact as I understand it is that tarmac driving techniques would in
fact theory get you through the rally course faster than a rally car. However, any error or unexpected surface or any other factor that is incorrectly judged for that matter results in loss of control and not finishing the course, which is not the desired result.
The general rally technique is to approach the limits, and then set the vehicle into a predicable oversteer and then release from the turn once pointed in the desired direction at a controllable speed. This technique is not as fast as the tarmac technique. However, when something does not go perfectly, the general mitigation is to hold the skid longer to reduce speed, or to adjust the rotation of the vehicle. Once predictably recovered, then start building the speed back up again. This technique is a bit slower than tarmac, but is much better at consistently handling the unexpected.
Of course, there are further nuances on timing when to enter the turn, when is optimal to complete the rotation, and planning for road curvatures of changing radius.