I remember reading somewhere that with proper exposure, compensation can also be used to improve local contrast in the mid-tones, especially for portraits.
The OP is proposing using a fixed number of agitations instead of a fixed time-interval.
My knowledge of chemistry is rudimentary, but this makes perfect sense to me.
If a chemical reaction only proceeds half as fast and thus takes twice as long, then doubling the time-interval (giving the same number of agitations) will produce the same result as double-speed and half-interval.
However, the rate of diffusion through gelatin does not depend on dev-rate, so that complicates things.
Anyway, I have created a concentrate giving XTOL-quality, and I designed it to have twice the dev-time as XTOL. And I recommend agitating every 60 seconds instead of Kodak's recommendation of every 30 seconds. It's worked fine in my experiments.
Has anyone mentioned "diffusion" yet? (Doremus used "infuse"...)
The reason for agitation has little to do with chemical "reactions" but a lot to do with helping the chemicals diffuse into the gelatin matrix of the film. The agitation ensures that fresh developer solution is brought into physical contact with the surface of the gelatin so that fresh developer gets into the gelatin and it helps remove development byproducts from the gelatin faster than if there was no agitation.
Still, in my book this means being able to get (squeeze) more usable information into the printable density range of the negative. It's really a question of semantics.
Just a couple of thoughts here:
First, the time between agitations is long enough to allow the developer to exhaust in areas of greater density (compensation) should be fairly constant for a given film/developer combination.
Now, if you want a compensating effect, you need to make sure the time between agitations is long enough. It would seem to me that this time would be longer with a more active developer than with a less active/weaker one. Therefore, to achieve a certain amount of compensation, one would have to increase the agitation intervals for the more active developer, not decrease them as the OP is suggesting.
Similarly, for weaker developers, the interval at which a certain amount of compensation occurs will be shorter. Therefore, for a given amount of compensation, one would have to increase the frequency of agitation compared to a more active developer, not decrease it.
Of course if you don't want and compensation, then just agitate away, it shouldn't make much difference at all as long as the agitations aren't long enough for developer to exhaust in the highlights; any agitation scheme that accomplishes this would yield the same results as long as development time was appropriately adjusted.