Ha! Bubba, I don't like any variation. However, for the most part that is unrealistic. For one thing consider this - your scale. How accurate and reliable is it? There is likely variation there as well. So, my plan has been this. For progressive loading, with a stick type powder, I try and stay will small kernel stuff. I'm not crazy about ball powders, however they have their place. Either of these will do well with the Dillon measure. For anything as large as say Varget, I think that the RCBS with the Sinclair baffle mod works ok. Not great, but acceptable for most stuff. Now, and this is the important part, what range and application is the load designed for? The longer the range, the greater the variation of velocity will impact vertical dispersion. Your standard deviation will give you a number to work with, and that is what I would suggest you use as your point of reference. Also, some loads hit a sweet spot and the variation has some range that keeps the load still good. Lastly, primers play a role. When I shoot groups for accuracy testing, I shoot a 10 round group, not the three round that tells little.