It's more than just powder size or brands of powder folks.
Part of the accuracy question is the consistency of the amount of powder that you use. Sure..., you are using a measure set or made to throw X amount of powder... but how consistent is that particular powder measure when using either granulation? What is the powder charge, and how much is the variation effecting the velocity?
So say you are shooting 90 grains of 2Fg. Your powder measure sometimes varies +/- 4 grains, so that's slightly more than +/- a 4% variation in powder, and the MV is probably very close to your optimal 90 grain load, so you may not notice.
OK you switch to 3Fg and reduce the load by 10%, which is 81 grains, but your measure settings are in tens, so you set it at 80 grains, but with the 3Fg your measure is only variable by +/- 2 grains. So that's a variation of +/- 2.5%... and again you probably won't notice.
NOW say you're shooting some light target loads, of say 50 grains of 2Fg. With the same measure variation as above, that's +/- 8%, a total range of 16% variation. While with the 3Fg..., the variation is only 4% for a total variation range of 8%... that's the same amount of variation you had with the 2Fg 90 grain load, and only half of the variation you currently getting with 2Fg in the lighter load. You probably will see the difference. Not to mention the fact that the slower burning of the 2Fg added to a powder charge that is off by say 4 grains may be considerably slower when using a light target load compared to the 3Fg.
The smaller the powder load the more the variations in what the powder measure throws is amplified, for the variations are a larger part of the whole.
LD