Go to the troubleshooter website in my signature and pull down the water troubleshooter. Let me know what you think.
Thanks. Got it. The thing is, I'm not measuring for a1, a2, and a3 (though I'm an intermittent lurker on the HBT science board so I'm familiar with the concept.) My method is a gross simplification. FWIW: I measure an actual DI pH (unless I have a COA which saves me the trouble) and where using multiple malts figure an aggregate. I then simply assume, as a starting point, a buffering capacity of 32 mEq/(pH*kg) for the typical mash (which I'm sure I picked up somewhere attributed to work by Kolbach and has long been stuck in my head.) Multiply through the desired shift in pH to get total mEq acid required, account for the total equivalent in (positive or negative residual alkalinity * water volume), and what's left is to be supplied by my acid addition. (Or the other way round in a dark beer requiring additional alkalinity.) As I said, this has been getting me right on target, probably within my margin for measurement. If the actual mash result was off on a particular malt, since I know DI and mash pH, and the total mEq acid or alkalinity, I would simply, in the future, assume a new, inferred buffering (never had to do this yet.) Technicianing it, not engineering or (Heaven help me) theoreticianing. It works. I'm curious just how horrified you are at such an approach. I'm just a practical brewer who wants to reliably predict mash pH, and this works far better than the available software (yours excepted, as I haven't tried it.) Why my quick and dirty method so outperforms professional software is the next mystery. Maybe some are overthinking?
EDIT
(Note: I've recently gone back over a large number of brew logs, checking actual mash results against what this old school method predicts, and what other software would predict, and just can't find a mash where that magic number of Kolbach's of 32 plugged into this method seems not to be confirmed very reliable by my experience, while results from other calculators are quite varied and unreliable. Douglas Adams was wrong, the answer isn't 42 after all. Whatever you can make of that.)