Did you keep info on boxed ammo accuracy or were you just burning bullets down the pipe?
There's an old saying, something like, "If you throw enough jump at the wall, one jump will stick out more than the others." Something like Seriously though, @rardoin is absolutely right, and that is the secret to the SMK success as one of the most accurate loads for the 308 ever devised by a manufacturer. I mean, 90% of the guns in the world can't be wrong, and 95% of the guys shooting Federal Match are not shooting "match" chambers with -0.005" jam, you know?I'd also suggest you drop your powder charge steps to 0.3 grains at a max. I prefer to go 0.2 once I've gotten to within a grain of my preceived/calculated maximum based on my knowledge of how the gun handles max charges with that particular bullet construction. You might find a window in there where you get all the resonant bunnies to jump together and point the pointy thing in the same direction every time it leaves the tube. Good luck!One more thing, are you sure you can't get those to load out to 2.820"? I get good reliability at that mag length with my Magpul, Lancer, and GI mags.
You might have a bad barrel, but probably not. My 18" BA is very accurate...if it shoots outside of 1 MOA it's usually my fault. I won't go so far as to say it's MOA all day, but I have made several .75 ish groups with factory loads. A few even better than that.It *definitely* prefers 140 eld-m and 143 eld-x factory ammo...but I haven't tried reloads yet do I can't really speak to that.I hope yours isn't a lemon, but if it is I'm sure BA would make it right.
@gnatmm A rule of thumb I use is: starting charges for cup and core bullets is pretty much interchangeable with one caveat - you must know you are not changing the internal capacity of the case more than 10%, and that you are not jamming the lands at your chosen OAL. Bullet construction, for the most part, allows for some liberties to be taken when substituting one mfg bullet for another. You cannot do the same with powder. Even swapping primers should cause you to back off your load data 3-5% and work back up to make sure you are not introducing a dangerous variable. Hope this helps.