In the formula desitter came up with, he is assuming that the water under the false bottom does not lose heat to the grains, and this is false. Even though it does not contact the grains directly (assuming that you add the water first and then the grains), the temperature will equalize in your mash tun.
While you're not wrong about the mixing, my formula does take this equalization into consideration, and my comment does make sense when you think about the problem a little differently.
Try this, two volumes, Va and Vb are the volume of the mash water above and below the false bottom respectively. If I was to just mix the volume Va, at some calculated temperature Ta, with the grain (at temp Tg), then the mixture would equalize at a temperature Tt (lets say my target mash temperature) which is a function of Ta and Tg. If I then mix this volume with the volume Vb which is also at temperature Tt, the temperature of the new mixture would remain at Tt, as I am neither adding hotter nor colder water. But I am still accounting for the volume beneath the false bottom. Hence "the volume under the false bottom does not lose heat to the grains" is correct in this context, as the above volume and grains have already equalized to the same temperature as the below volume.
The formula uses this reasoning, but instead determines the strike water temperature required as a single infusion, rather than the multiple infusions at different temperatures given in the example above. It then uses this strike temperature to derive a new strike water temperature that also allows for heat lose due to the volume of your mash vessel in contact with the mash (if anything needs adjusting for accuracy due to heat dissipation it's this, but for simplicity sake, and to do it just as BeerSmith does, I left it like this).
The main concern I have, and I think many others do too, is that we are not getting our desired water-to-grain ratio due to part of the strike water volume being below the false bottom, and hence the mash becomes thicker. Although minimal changes in the water-to-grain ratio likely have little effect, I'm still concerned by the lack of consistency when varying the amount of grain. It may be true what you say with regards to the higher water-to-grain ratio mash efficiency increase offsetting the lower efficiency due to less sparging with larger grain bills, but I think the software should at least attempt to model what is really happening, and strive for accuracy if at all possible, especially in this case when it is a relatively simple addition to make.