SAA, Actually I do not think that the formulas of powder have been tweaked-EVERYtime I have checked new lots of powder against old lots of the same name they have been extremely close in performance.... but my new lots have actually acted very slightly slower.....so they would in theory permit fractionally higher charges. I do believe that there may have been wider lot to lot variations at times than we might wish to imagine. Mostly I think that the more "energetic" loadings -particularly of the past reflect two things. One is an attitude difference that is willing to push things closer to the limits of safety and the other is that low pressure cartridges are actually hard to pressure test consistently......Furthermore,most current data is conservative ( I am not saying artificially kept low just cautious in seeking to keep all individual readings under the max average allowable) and SAAMI standards have varied some over the years.
Having said that , I am not willing to shoot Speer #8 38 special 4756 loads in my 38s......most of the data with other powders is "warm" but not truly excessive having been loaded for years , is comparable with other manuals, and never deleted from later editions of Speer 8.
OTOH current 38spl and .357 mag data for 4756 makes no sense whatsoever-it doesn't reflect anywhere near actual velocities with any lot of the powder I have ever used, and is very incongruent with the relative burning rate of the powder. I speculate that the team on Speer 8 and its contemporary Sierra crew had a lot of 4756 that was a little slower than any other lot and that shortly after publication Dupont released the fastest ever lot of 4756.......and problems surfaced. I suspect that Dupont, then IMR, then Hodgdon have submitted batches of the overly fast lot for testing ever since trying to avoid a recurrence of the dangerous problem.