There's an axiom in precision shooting, "the name of the game is the same" referring to the belief that eliminating variables leads to more consistency. I'm not arguing that point, but it often ignores the biggest variable in my group sizes: me as a shooter. I try to control what you can, but I don't want to spend so much time chasing hardware variables that I forget to focus on the shooter. Practice more. Dryfire more.
I started casting using a borrowed .490 mould and a cast-iron ladle heated on a gas stove. I got occasional fliers--probably me, but I always wondered. Over the years I got curious about possible flaws affecting accuracy. For about a decade I had a cast-bullet business (didn't make much money but I learned a lot). Different people get different results with different techniques; part of this is probably equipment/alloy related. I'm a bit CDO about casting (CDO is like OCD, but with the letters in the proper order as they should be). What works for me may or may not work for anyone else. That's life.
When doing experiments, breaking in a mould, etc, I use lead from Rotometals to avoid unknown contaminants (while their lead isn't "pure", 99.9% lead is closer to "pure" lead and more consistent than I can salvage from scrap). I use an IR thermometer to track temps of the mould and lead pot, and a metronome to get my timing right. I pre-heat the mould on an electric hot pad to get to my starting point, and sometimes use a fan or wet sponge to cool it down when it gets too hot.
I tend to weigh everything I cast, but usually find little variance. Pretty consistently, after throwing out the first few balls cast, without taking particular care after a visual inspection I get a range of -.3gr to +.2 gr. This is with small stuff like .457s or .535s or even .735s. With care, I can usually cast a run of 100+ 45 caliber bullets for the long-range ML with a range of .1gr (i.e. plus-or-minus .05gr). But there's a big qualifier here:
I'm not sure it matters. I've done experiments with long-range round ball shooting (200/300/400 yards) and seen no statistical difference in sizes of multiple 10-shot groups using balls with a .1gr range/a .5gr range/1.5gr range. At closer range (50 yards) using a heavy scoped test rifle and it's best load fired from sandbags on a bench, I got a >5% difference in group sizes when shooting balls sorted into the same .1gr range/.5gr range/1.5gr range.
With handguns I did my tests using a Ransom Rest mounted on a concrete base. My test gun was a Ruger Old Army shooting .457s. I found no statistical difference in group sizes at 50 yards between the .1gr/.5gr balls, and minor differences (possibly not significant if I had compared enough groups) with the 1.5gr balls.
The biggest effect of eliminating possible defects was that it gave me confidence in what I was shooting (it also took away the "flaw in the ball" excuse for fliers, but you can't have everything).
