I have found in my somewhat limited (about 6 months) experiments using dilute Rodinal with infrequent agitation techniques, that a notable deal of "speed" goes down the toilet if you go over 1:100. If you are doing this, I'd overexpose as least one stop, and maybe more like one and a half or two would be better. I have been doing T-Max 100 with 1:100 semi-stand (5 minute intervals until a certain total agitation time has been reached), overexposing two stops (rating at 25 as a cheater way to apply across-the-board EC to the whole roll). Results have been very unique. Going to 1:200 didn't seem to do anything except kill even more low-toned detail and make the grain too coarse for what I wanted. I hear from Ian Grant here on A.P.U.G. that T-Max, if measured by Ilford's method, would end up being a 50 speed film, so if this is true, I am only overexposing one stop.
P.S. If you want to pull a good deal (more than three grades), but still want to maintain local contrast in shaded/dark/overcast areas, it is the best way out there, IMHO. Just be sure to use a very fine-grained film if you are not a fan of grain, because it really does jack up the grain a lot IME. I most often use this technique in: 1. night photography with lights (Neopan 100 Acros or T-Max 100), 2. landscape photography in which a large and detailed part of the shot is in shade, but "shafts" of sunny 16 light are also shining through illuminating certain parts of the scene (Pan F, T-Max 100, Efke 25, Rollei 25, Efke IR 820C, Rollei IR), 3. severely backlit shots in which I want detail in the foreground without the blocked up hightones "bleeding" into it, e.g. a sunset or an otherwise super-bright sky shot through a tree.