4 cells will stay in regulation longer as I believe the Malkoff Quad uses 4 separate buck converters, one on each emitter. Buck converters only work in the input voltage is somewhat higher than the output voltage. Under heavy load, 3 cells would definitely fall below the output voltage and thus fall out of regulation while they still have capacity left. 4 Cells should remain in regulation the entire time.
Also, I'd want to have a copper pipe as a spacer, not PVC, with flashlight temps approaching 150F.
Ideally, the copper would actually make physical contact with the quad, thereby creating an additional thermal pathway away from the head, aside from the poor path through the threads, and down the skin of the light.
There are two ways this could work:
1) Use a cutdown mag body with a tail switch, and let the copper pipe run all the way up to the quad.
2) Remove the stock switch, then cut a hole in the copper pipe and install a Mag C Switch INSIDE the pipe.
A Cutdown ~1.5D mag running 2xAW C cells, with the copper lining inside as I described would be a pretty nice host for the quad. Nice and compact, still decent runtime, and it shouldn't reach excessive temperature.