I think it's possible that off-road braking is significantly different than road braking. In off-road braking, you could easily exceed the traction of the front wheel and cause it to start skidding; at that point you've certainly lost your most effective braking; sliding friction is a fraction of static friction. With a loose surface, you absolutely want to use both brakes and de-emphasize the front brake, because you need square inches on the ground to provide traction to brake with.

The same would be true on any surface where the front wheel's traction was not sufficient when braking to lift the rear wheel from the ground; gravel roads, wet pavement, snow, ice, etc. In fact, on ice and loose gravel or sand, it can be very dangerous to use the front brake at all. However, once you get to the point where you've got so much traction on the front wheel that you can flip the bike over it, I don't really get how the physics could support anything but maximum braking on the front wheel, just to the point where there's hardly any contact on the rear wheel. If I'm braking 90% front, and 10% rear, and I increase the front braking power a bit so it's 100%/0%, and I still have not started skidding the front tire, I don't believe that I'm slowing down SLOWER than I was before I increased the braking power on the front.

It certainly is not a good idea to go all the way 100%/0% in almost any situation though, since you'd have almost no control left. But if I'm anywhere near that, I'm going to let go my rear brake, because the rear wheel would skid with almost any brake pressure when that much weight had shifted forwards, and I do NOT want either of my wheels skidding in emergency situations; skidding = loss of control and much less traction.

I am not insisting that I'm right, but I can't visualize how the physics would support the partial-rear-brake balance as stopping faster. I'd like to see the study, or if someone wants to draw up the force vectors of the two situations, I'd be very interested.