You will never get the case shoulder between the “Go” and “No Go” by following the typical “size to the shell holder and add a quarter turn” instructions found with sizing dies. You must use a case gage to verify properly sized length.
"Never" is a mighty long time. I guess it has expired though since I've yet to have a rifle fail to go to battery on one of my .223 reloads (shooting them in 4 different .223 and 5.56 ARs) and I don't own a small base die. If your rifle needs one, fine, but most don't.
Never is a long time, but it always takes less time to type “never” than go into some long explanation over gaging, tolerance stackup, probability and randomness.
You have not said whether you use case gages to set up your dies, but I am going assume you don’t.
There is enough inertia force in a AR bolt carrier to crush fit a slightly over long brass case. I think steel cases is another matter. Bolt gun cocking cams are powerful enough to crush fit a case to the chamber, but these cases can only be a little long, or little fat.
There are threads where people, following the die instructions, can’t get the bolt closed because the case is too long.
Mike Venturino, the gunwriter, is one of those who for years has stated in print, that small base sizing dies are not needed for gas guns. I find it humorous that in the July 2012 of Guns his reloads were so fat and so long, that with the AR10 and M1a he was testing, he could not get the bolt closed. He could not fire the rounds out because he could not get the bolts into battery, and the rounds were so jammed in, he had to take these rifles out back, with live rounds in the chamber, and beat the bolts open with scrap lumber, to remove the cases.
Ha, ha, poetic justice.