It's off topic I guess, but I've done a lot of testing on Microdol and Perceptol at various dilutions with various agitation routines. When the curves are plotted, even at 1+3 compensating action is minimal. As for speed, my advice to anyone wanting to fully exploit the image characteristics of these D-23-based extra fine grain developers has always been the same, which is to determine your working EI at stock strength, and use that same EI for 1+1 and 1+3 (ie accept the speed loss regardless of dilution). I've found graininess increases very quickly if you try to get more speed - even at 1+3, and it doesn't take much until you are gaining nothing grain-wise versus D-76, XTOL, and other developers that will give you better speed. This behaviour is consistent with a developer that is not truly compensating, and is weakly alkaline.

The other point is that I still wouldn't consider OP's Tri-X "badly" overexposed. We're talking 2 stops over ISO, probably closer to 1 stop over the working EIs people often use. And since this was a low contrast scene, there is plenty of room on the curve to accomodate the luminance range. OP could use Perceptol, but I don't think there is anything to be gained in this particular situation vs D-76. As for pulling, that will decrease local contrast in an already low contrast scene. A compensating developer would also flatten local contrast.