Hi Rich, freddd's protocol reportedly has been developed by laborious trial and error There was a lot of targetted trial and only small amounts of error. I did a lot of reading before adding anything new and had good reason for trying each addtional thing. The glutathione precursors were the only real error in six years. Except that not all brands are equal, not all injection solution preparations are the same, even from the same compounding pharmacy. I don't know of any theory that would weed them out in advance or even expect them. There was a lot of debugging; finding why "X" didn't work and correcting it until it did work. That was laborious. Systematic testing of injections from 1mg to 25mg until finding the shape of the response curve, the color of urine, correspondence with sublingual doses, responses and urine color. A new one now is to correlate the dropoff of effectiveness to minutes of light expsosure at a certain intensity. As some have said disparangingly, "That's just engineering". However, the "just engineering" is why the system works quite predictably and reliably. If software was devoloped the way medications typically are we would have software that worked "a statistically significant percentage of the time for a statistically significant percentage of the people. The IRS would love it when you explained that you are one of the 33% for whom accounting software doesn't work. And being in the 1/3 in which it doesn't work aall is the pits.