Hi,
The OP's question can be at least partly answered with what will initially seem like a flippant reply: the boys in the white lab coats with millions of dollars worth of test equipment know a he(ck)uvalot more about loading ammo than any of us do out here, in that they can actually test stuff and put real numbers to it, not just comment that "it ain't blow'd up yet sos it must be safe."
But please bear with me. There are lots of reasons loads are rejected or seem to be ok but aren't included in the books, and they generally come from those lab results. Let me repeat a story I've told before about my uncle:
His working life was devoted to making and testing solid rocket fuel, "propellant" in pretty much the same sense smokeless powder is. In fact, they started their test sequences on a tiny scale, loading shotshells! From there, they developed tons of data and computer models (this was a multi-billion dollar company, so you know they had the wherewithal to have the finest equipment at their disposal.)
Well... one day he was going to test a load that, from all their previous testing and modeling, looked like it would be perfectly safe. Things had been nearly linear to date as far as results and there was no reason to suspect that would change yet. In terms of this discussion, it would be like adding a few tenths of a grain of Unique to work up from just under 12,000 psi, which had been tested, trying to get closer to the SAAMI max of around 14,000 psi or so which was about to be tested. Everything was prepared and carefully checked.
The firing of the shells was done in an enclosed device that I understood to look like a blasting cabinet from the outside. My uncle placed one of the shells in there, and touched it off. What happened next nobody could explain, except that experience told them could and does happen: there was an unpredicted pressure spike which blew the test fixture apart, injuring him, and he nearly lost his eyesight. It was touch and go with his eyes for weeks.
This is just one small example of the kinds of things the ballisticians deal with daily. Sometimes safety is the problem, as it was in my uncle's situation. Other times, accuracy, reliability, consistency, and/or other non-safety related performance issues may play into the decision. Environmental conditions can be a factor, too: when I was shooting trap competitively, I had a favorite load that worked just fine, in balmy to hot temps. Just above freezing, though, it became a whole different thing, what with the "pop, bang, boom, pop, bang" situation. I've also had one particular rifle load that was fine from freezing to about 90 degrees or so. Somewhere between 90 and 100, it became what I'd call uncomfortably if not outright unsafely over pressure. I'm sure many of you here have experienced similar situations and have your own stories.
Point being, there are tons of variables we don't know about, and the ballisticians aren't likely to let us in on, which show up in their testing. So they don't include the load. Yet one might find the very load they're looking for, which was not included in Book A, staring right at them from the pages of Book B. The guys writing Book B simply used different parameters than the guys writing Book A! Thus the ever so common advice to get multiple books and compare/contrast between them. "Good" loads tend to show up being remarkably the same across several books, while that load shown only in one of them might just be suspect.
In the end, we're responsible for the ammo we put together, so a bit of extra caution's not a bad thing... stay safe!
Rick C