I have a current spreadsheet open, which looks at the issue of an estimate for 'best' firepower as a product of frontal area, penetration at a pointed range and the hitting space of a nominal target nearer and further from the nominal range from the lower of apex and top of target to the bottom.

Coupled with an approximation to internal ballistics after Helie and external ballistics after Dr AR Collins/NACA/NASA this allows a trace of 'expected' 'best' hitting power for guns in distant fire which falls off with additional elevation from the range for a gun pointed by the line of metal and distant charge... but which holds this pointing and reduces charge ratio for standard, reduce and the three 'distinct' velocities from the two lower charges with double shot (the middle value is approximately the same for the bottom of std/dbl and the top of reduce/dbl).

The resulting curve is a 'flat topped' pointed gun, and a falling off in random fire, with the 'corner' range both modifying short range hitting power and the 'best' range for direct pointing - before random fires, sighted fires or fire in ricochet is needed to extend the distance.

Carronades and other single charge sighted pieces just have a single curve below but 'similar' to an equal calibre gun in random fire. The carronade shares a common performance with the gun's 'common' double range and hitting condition.

Larger guns will always hit harder within the random fires area (both more powerful guns of the same calibre, but more strongly a gun of larger calibre). This combines so that the 'best' performance of a short 12pdr is worse than that of a 32pdr carronade, near to the range of direct pointing of the gun, and is much worse at very short or very long ranges.

An unintuitive result is that guns pointing to a longer range (steeper taper, usually where breechring and muzzle swell are similar across distinct lengths) are often shorter, weaker pieces than others in their common calibre and the RN guns tend to a lower pointing than French canon, especially in larger calibres. The 'flat' of a 36 Livre gun is similar to a 24 pdr gun despite it being considerably more dangerous in random fires and having a ~800/600 range proportion with the same input parameters and the appropriate geometry. The potential range of pointing by LOM without the use of a hausse/tangent sight but only relying on the divisions of a stepped/notched dispart for a 32pdr carronade is then found to be ~910 in proportion, although a carronade is always stronger at shorter ranges, as well as the pointing at a larger apparent target form, and an easier time with the estimation ranges which is common to all ordnance.

A weaker or stronger powder is similar, but not identical to recasting the whole to a shorter range without materially altering the strength of the pointed 'short range fires'.

The quoted maximum effective range of the 95cwt 68pdr allows a similar accuracy or similar expected hitting power to be used to consider the 'similarly' effective firing ranges of guns of various patterns and sizes as well as the carronades, obusier de vaisseau and shell-guns. Combining the proper number of pieces of each calibre and type within a class/establishment/variation can then allow an estimate for the potential firepower of each considered platform across the whole range of engagement, and shows French vessels to on average be stronger at longer ranges, but to be weaker in the closer fight, mostly because of the shape of the long-guns.