Both my .45 and .40 cal rifles show better accuracy with the larger balls over smaller ones. The examples I've tested are .395" and .400" in the .40cal., and .440" and .445" in the .45. In both rifles, the larger balls are more accurate.
in the .69, a 15 bore ball (.677") with very heavy .030" denim patch (that is completely reusable for exactly the same accuracy for several subsequent shots) is NOT as accurate as a .684" ball with either a .022" denim patch or the same .030" denim patch. The heavy patch is the most accurate load, better than anything looser.
My .58 Enfield shoots just fine with a .562" ball and .020" denim patch. The bore is a tight .574", so that ball is .012" smaller than the bore it's going in to. The rifling is only .003" deep at the muzzle, then gets deeper towards the breech, which is .012" deep. The bore remains .574" for it's entire length - only the rifling depth changes (same as original US military Minnie rifles of the 1850's & .60's). This short 24" bl. rifle shoots more accurately with a .575" pure lead ball and a .018" denim patch, but of course, getting the ball flush with the muzzle takes a heavier blow of the starter's knob. Ones hand must be fairly tough or it gets sore. Using the smaller, looser .562" ball allows me to use WW metal with the same .020" denim patch.
So - all guns have their preferences, BUT - almost without exception, they'll shoot best with a tighter combination. Exceptions might be out there, but it takes a lot of testing and at least a couple of pounds of powder to find a rifle's 'best' accuracy, not only what's 'good enough' for the persieved job at hand. The bigger the bore, the more pounds of powder it takes to test all variables.
Many guys pick a descent patch, then a ball mould that's easily aquired, then merely adjust the loads to find the best accuracy with that powder #, that patch and that ball diameter. That's only testing one variable. The ball size is another variable, the patch thickness is another variable and the lube is another variable and the powder # is another variable - not to mention different powder makes, which are not available here. You guys down South have the last variable - lucky- or not depends on how you look at testing rifles.
My last range test was with the .32 Ten. rifle & I accomplished one variable only - the powder charge - that took over 75 shots with not enough rounds per charge to test for consistency. I need to now double and triple check my previous results, let alone testing the other variables. Testing is only as complicated as we make it, or allow it to be, but it also makes a big difference in how competitive we're going to be - or perhaps want to be.
A poorer shot like myself needs an exceptionally accurate load to be competitive. Guys like Taylor don't need load development over the standard recipes we've developed over the years, loads that shoot reasonably well in all guns. Guys like Taylor, Mark and Neil can shoot offhand almost as well as the load they're shooting will print off the bags while I need all my ducks in line to best him/them and sometimes, only sometimes, it works.
Sure makes me feel good when it does. Makes it all worth while.