I just reread the study I linked to in post 13, and he did find a 1-4% variation in pressure in the two rifle cartridges (223 and 308) he tested. The SD of SAAMI MAP for rifles which is the basis of the stats for their system, is 4%, and the SAAMI MEV (maximum extreme variation) lets individual rounds within a sample go to a theoretical +18% (high but well below proof level). So I will amend what I said in post 25 to say shooters without pressure measuring gear saw no apparent difference in pressure that I am aware of, and the difference of up to 4% will not cause overt overpressure signs if you are loading within the normal data range. However, you can eliminate even that 4% difference by reducing powder charges by about 2%.
I noted on the re-read that the powder used in the paper is WC749, which is a spherical propellant from the St. Marks, Florida plant that GD owns. 748 is available in canister grade in the Winchester brand, but 749 is not. And while 749 is probably close to 748, I don't have any independent information on that, so it remains to be confirmed. Being harder to ignite, the older St. Marks spherical formulations may be more affected by flash hole size than some other powders are. In past experiments I did with flash hole deburring and uniforming, a spherical propellant (2520) was affected measurably by flash hole modification while single-base stick powder didn't seem to change behavior at all. So I don't know how universal that 1-4% effect may or may not be.