Well, while Thornton's heart may have been in the right place, when I tested a lot of these things I found the chemistry logic to be flawed, or at the very least, overcomplicated for no reason. I'm picking on Thornton here because that was the topic, but it's really two-solution Metol-sulfite-alkali development in general where there is much confusion and conflicting information on everything from which alkali to use, to agitation etc, etc. As Grant Haist points out, scientific studies on true two-bath development are scarce, and the situation is even more nebulous when it comes to two-solution Metol-sulfite-alkali development. I've seen virtually no good data, testing, or even sound theory for what most people say.
Suffice it to say before changing the formula, I would first start by reducing the amount of time in solution A. You'll have to experiment to see what gets you the right contrast/speed balance.
Also, another flaw I've found in the vague instructions regarding solution B in general, is that frequently the time is too short, and the effects of agitation are not well documented. This depends on the pH and buffering of solution B, as well as the film type.
My view on this type of development from a sensitometric perspective is this: the aim in most cases should be to develop is little as possible in solution A, but for enough time that the emulsion becomes saturated with developer, and then to develop in solution B until exhaustion. This maximizes the benefits of this type of process: (a) a long, straight curve, (b) lower than normal contrast, (c) retention of film speed.