If you check the ranges in M&B, you will find that there is almost never a shot at more than 50 paces. Do some tests. Have your little guy walk across an expanse in town, and count the paces. Or measure in lengths of a horse. It comes out that shots you thought were 50 paces are actually 10. It's a graphics bug caused by the fact computer screens are not as big as reality.
(On the Holy War mod, trying to get arrows to shoot right, I guessed a shot at a merchant sign to be 25 paces. When I checked it against the size of a horse, it was *8*.)
Also, mis-sized shot is not a valid complaint. With black powder, you wrap the ball in wadding when you load it. With two extra patches, you can fire .45 cal ball from a .54 rifle and still get pretty good accuracy at short range. (It increases your pattern size by maybe a foot at 100 paces - I've tried it.) That's the worst shot mismatch I can think of, like trying to load .38 special into a .44 mag - and even then it shoots pretty good. Only modern bullets, which lack soft wadding, have to be precision made. The old-timers had this figured out too.
Even 99% accuracy on this game won't keep all your shot on a body at 50 paces, even with high skill. Most guys I know who play with black powder shoot better than that, even with primitive weapons like smoothbore flintlocks. (They have all the same problems of the matchlock, except that you don't have to keep your fuse lit.) And the Spanish arquebus I have seen from this period were relatively long-barrel weapons with fair-to-decent sights - based on having fired a lot of ammo through muzzle-load black powder weapons, I figured they would shoot pretty good. There is no "sniper weapon" on M&B, because you can never see more than 100 paces or so.
Lead round ball from arquebus should sometimes deflect from steel armors. Armorers of the period would "proof" armor by firing a musket into a corner, usually the lower right hand corner of the breastplate, to prove that they worked. Many surviving armors of this period contain these "proof marks". It's not a bug. It's just that bullets don't go through everything like magic. (My tests with modern handguns say this is not rare - modern pistol rounds won't go through much metal either. High-velocity rifle bullets will - but the arquebus was not one of those.) The armor and bullet penetration numbers are pretty accurate. Test them if you don't believe me.
I'll stand by the horse numbers for two reasons. One, they are accurate. Horses do that - they're not cars that have to turn a wheel and roll. They spin on their feet the same way you do... do you need a large turn radius when running, or do you dig in your heels and basically run right back up your own tracks? Hard to stay on one when he does that, but we can't simulate that. Two, if they are lower, horses will be useless - they already get torn to shreds by the arrows. They need all the advantages they can get, or else they are just target practice for the enemy.
The sword damage thing was done quickly. Put a one or two point variation between them if you want. Don't make it more than a couple of points. That will deliver the message that they are a little different, without changing the basic function.
And the foil did exist, just that nobody considered it a weapon. The closest thing to a weapon foil was the estoc - which is present. You'll notice I removed the cutting edge from the estoc - they are square cross-section, and can be used as a minor club but have no cutting edge.
And as for the question about melee weapons piercing steel ... test some hatchets on several layers of auto-body weight steel. See how many layers a cheap hatchet will go through - it's pretty amazing. They penetrate more metal than many modern handguns. A flint axe (despite being heavier and a little clumsy) will perform almost as well as a steel hatchet - I've seen it.
I say the numbers are generally good. (Put some variation in the swords if you want - every sword is a little different.) The armor and missile weapons values will stand up to tests. And check that first link that LCJr posted - it shows a lot of what I said (the reasonable accuracy at short range, even from the hip with no practice, and the mediocre armor penetration), even though the weapon is much older than the ones we are simulating and had no way to aim it.