I think part of the issue this game has with accuracy is comming from two major issues:
1. The point of origin is not on musket/rifle but actualy variably on at least 3 different location depending on FPS/TPS perspective and even if you are crouching, which put the origin weirdly more behind your back. This lead to not only potentional exploits as addresed in my post, but because the projectile's path starts actualy above the player, close range accuracy is all over the place.
Also given that the point of origin start behind the character and atributes a random horizontal spread immediatly at that point, you can have litteraly point blank missed even if litteraly stuck your weapon's muzzle into the enemy as for now, the muzzle and basicly the whole rifle/musket is nothing but a decorative object.
2. Randomly selected values when calculating the trajectory. While variation in the muzzle velocity seems like a neat idea to simulate potential loss of gunpowder when loading the musket/rifle, the values that are picked are not sensible and are overexagerating a marginal loss of powder. The random gravity values is just bad physics.
What is kinda sad is that there isnt a spread as would be expected around a predicatable flightpath, but first, you have from the get-go a random horizontal spread, then you have a vertical spread determined by the random pick of the muzzle velocity and gravity values. And while on the rifle the vertical spread is not as severe given that it picks values that are close to each other, the horizontal spread seems to be determined by a random pick of +-20 degrees of angle immediatly at the point of origin. Returning to a previously posted image, you can see that the vertical spread is minimal, while the horizontal is totaly random nonsense. (note that if this test would be conducted with a musket, the vertical spread would potentionaly be also all over the place)
Also as an oberservation. Given that the flighpath originate in the crosshair, this prevent from shooting when you are using the free-look key as the crosshair is not present at that moment.
To address the point why some peope including me are now vocal regarding the awfull handling of the shooting mechanic is that this game had closed testing for a while, with people I have known from playing NW DLC for Warband. The shooting mechanics seems to be the result of these closed testings and given that developers commented on fixing and improving melee while there were no mentions of improving the shooting mechanics (I may have missed them, please correct me if I did) I have a bad feeling that this is because everyone was saying "this is fine".
There is also a common argument I have noticed around, that it is to simulate inexperience. I will not address this point directly, but by examinating this cherry picking. If it is realy to simulate an inexperienced shooter what about melee then, melee training during that period was marginal compared to shooting, should we be happy with very bad melee implementation as it would now simulate the "lack of training"? I think no. I think that all mechanics in this game should be refined.
Leon, you cannot argue that there are no linebattles destroyed by this accuracy problem when there were hardly any linebattles conducted since the game's launch few days ago. I was invited to a linebattle but couldnt participate, but when I questioned some of the involved players the feeling they had from it was realy lackluster. Other players I know which are critical of many issues of the NW DLC for Warband said, that holdfast made them appreciate the NW DLC more and that they are beyond the 2 hours refund point anyway.
To get back to the implementation of firearms, these things are I think woth consideration.
1) Trajectory originate's in the musket (even if this would be the only thing implemented, it could improve the close range accuracy)
2) The trajectory calculation would use more consistant muzzle velocity numbers. The tighter barrel of the rifle would allow a higher muzzle velocity for the rifle leading to a longer, flatter trajectory.
3) Spread would be applied as a randomly picked degree of angle on a 360° rotation around the bore axis. A musket would have a bigger number, while a rifle's one would be smaller.
4) Weapons would be sighted to a certain distance, making marksmanship more reasonable.
5) As a potentional change to the gameplay making realoading while moving slower.
6) Ammunition selection for riflemans, where they could choose the amunition type they are currently firing and decide between loose ammunition balls for quicker reloading to use for harasment or tight ammunition balls for a precise long range fire with the downside of a longer reload.
Also for people questioning 18th century musket making, this is an old, hour long, but interesting document which shows a gunsmith using period tools and techniques to make a musket.