TL;DR. I installed an MT probe under half way up the mash volume to compare with my HEX out temp to have confidence heat was being transferred to the grain and the mash temp was uniform throughout.
This is an interesting question. There are different approaches.
Some people install the probe so that it reaches into the 'centre' of the grain mass. I think this is reasonable because it helps assume a reasonably accurate 'homogenous' temperature for the mash. But there will be a gradient top to bottom depending on vessel insulation levels. Arguably, if you have perfect insulation in your vessel and you've stirred your water and grist so its temp is homogenous then your probe could be installed anywhere - in an ideal scenario that is.....but not real world. You'll recirc with HERMS, so volumetric flow rate matters then for HEX outlet only scenarios. If the flow is minimal because your grain bed is a little tight then it may no longer be a reliable indicator.
Also, temperature losses between HEX out and MT in need to be known. On the flow rate issue, your MT geometry may come into play. If it is uninsulated and tall, then without enough flow rate evenly distributed over the mash you could loose more joules of heat off the vessel skin than you would want, possibly meaning your mash temp is within your target for only the top part of the grain.
I installed a MT probe under half way up my average mash volume (the water/grist volume - not vessel) so I know what the delta is between my HEX outlet and how that impacts on the mash. Good for diagnosing any channeling you may have. Calibrate both sensors in that scenario. without the MT probe I'd be watching the HERMS element to see that it wasn't busting a gut trying to heat cold wort from the bottom of your MT.
A problem I've encountered with MT probe is that the thermowell gets in the way of my falst bottom removal. a slight hassel so keep that in mind for whatever cleaning plan you may have.
