Interesting thread.
Well, here goes.
The approach in MpH 4.2 is explicitly based on charge conservation as expounded by AJdL. If you look "under the hood" you will see that the calculations related to mash pH are entirely different than in previous versions (which all used the same math, IIRC).
Calculations that didn't change include equations that convert the amount of an added salt to (i) ppm and (ii) normality of the dissolved ions, but these calculations are straightforward and have no room for interpretation (aside from the hydration states of the Ca and Mg salts, which one can perhaps argue about).
Calculations that convert the amount of acid added to acidity (or, equivalently, alkalinity) also did not change. These, too, are are straightforward and have no room for interpretation.
The major complaint about earlier versions of MpH had to do with poor predictions of mash pH upon the addition of acid to the mash. This flaw has nothing to do with the acid-amount-to-alkalinity calculations, but has everything to do with a poor calculation of malt buffering capacity. (Indeed, I, too, see this issue when I run actual brewed beers through older versions of MpH). I will say that the calcs used in older versions of MpH were based on what I could glean from Kai Troester's experiments, which was about all the relevant information that was readily available back when I first developed MpH. (Aside, KT's data was pretty much what MB had when developing BW. Differences between those older versions of BW and MpH can largely be attributed to slightly different interpretations KT's data.)
The reason I put out a request for people to send me actual pH measurements from their brewed beers was so that I could find the best "universal" correction factor to go from laboratory measurements of buffering capacity on pulverized grains to the buffering capacity of crushed grains in a typical homebrewer setting. This is the "malt buffering correction factor" found on the "Mash pH v4" brewing tab. It very well may be that this factor requires slight adjustment for a particular brewing setup, but I would guess that 0.5 and 0.8 would be the extreme ranges for any situation.
For the beers that I have run through the various calculators, I find MME, BF, and MpH to be the most accurate. Now I've not quite yet run all the data I have through all of these calculators. In particular, I have yet to look at all the data you sent a while back. Time to do it, I guess. I will say, though, that I have run your batch 128 (Boont Amber) through MpH 4.2 and I get a difference of 0.01 [with a correction factor of 0.65 (which on average appears to give the most accurate results)], while you calculated 0.05. Obviously, we entered at least one thing differently, but this does not necessarily imply user error on either your part or my part. It likely is due to a difference in choice of a malt of two to use for the calculations. Perhaps you could email me a few copies of MpH with several of your beers entered? I would like to track down the difference, in case there is something systematic.
Cheers!