Finned ammo isn't that uncommon.
I could create a finned round that HEAL
could just about fire over 500 yards at an elevation of only about 5 degrees (and well over a mile at 30 degrees elevation), but there are not many "ranges" that long I wouldn't be seriously concerned about sending a chunk of pointy high sectional density steel down at half the speed of sound though, in case someone should end up in the line of fire by mistake.
HEAL lies in a quite rare class. Most low bore launchers (and most "sniper" launchers are low bore) would have difficulty in imparting as much brute force and kinetic energy into a round as HEAL can (simply because most don't use 320 psi, or in many cases, as long a barrel), so would have difficulty propelling a round out to 500 yards with a low launch angle.
But sub-MOA accuracy from a homemade launcher (most of them smoothbore) at those ranges is exceptionally unlikely. Even things like the Barrett M82 are expected to shoot not much tighter than about 2 MOA. The HK PSG 1 and the Barrett M99 can achieve 1 MOA or better, but those are high grade military rifles using match grade ammunition. (World guns is useful!)
Our standards are unlikely to ever rise above those of the military, so we have to set our sights a bit lower (there's probably a pun there).
Just because I can, I assessed the chances of hitting a 2" circle at 100 feet. For a 90% chance of a hit on any given shot, the accuracy must be at worst equivalent to 5.4 MOA. For a 95% chance, that falls to 5.2 MOA. 100% chance, no less than 5.1 MOA.
That's impressive for a spudgun, but that represents only 4% of the consistency of a 1 MOA rifle.
A 25 fold increase over that already very good figure, as well as requiring 500 yards is a nigh on impossible challenge.
So, rather than setting a test that would require a very powerful launcher to even reach the range, I suggest using this easy method to find your launchers accuracy in MOA.
The Ragnarok accuracy assessment:
Set up a round diameter target more than 10 yards away, then fire at least 10 rounds at it. The more the better.
Count the number of hits - shots that lightly clip the target are not counted, solid clips are are.
Then use this equation, remembering to use the same units for target diameter and distance to target - inches, centimetres, metres, angstroms - it doesn't matter:
The answer is given in MOA.
It goes without saying that the more shots you take, this figure will become closer to the actual value of your launcher's accuracy.
Use a sensibly sized and positioned target. If you are getting 100% hits, your target is too big or too close - you are shooting better than the test can actually determine a figure for! If you want an accurate figure, use a sensibly sized target at a range where you are getting about 50% hits.
You could use an Archery style target, and count only the hits within enough rings that about half the shots are within them, using the diameter of the widest ring you are counting as your target diameter and number of hits (distance to target and number of shots will be unchanged of course)
However, that's got me interested. If I can find a suitably large range to fire down which no-one will wander across, I'll make 5 or 6 finned rounds, screw on the porting attachment, set up a large paper target and quantify the accuracy.
I used to be able to shoot 2 to 3 MOA at Cadets if they let me key the sights in properly first, almost as good as the rifles could actually do.
If I can find that 500 yard range you mentioned, and HEAL will shoot straight enough, well... we'll just see.
EDIT: Made mistake with equation. Corrected.