I wonder how much those VRM temperatures you've measured with that probe correlate with the real tempeature of MOSFETs. If I were to make such a test of waterblocks I would get a pyrometer with laser pointer and measure the temperature on the back of PCB right under one of MOSFETs. It is also possible to measure the temperature of memory chips. Such a measure would strictly correlate with the real temperature of the component on the face side of PCB.