Hm...

The scarceness of the ammo isn't always a guide to the effectiveness of a weapon. Not nearly. 7.62mm ammo is common, whereas 5.56mm is less so (on a global scale); the larger ammo is more effective. There are lots of technological and socioeconomic factors. Not exactly what you want to hear, hm?

A society will build cheap weaponry that's effective for its military. The underground will use the same, perhaps in slightly outdated forms, or else stick to smaller weapons, often similar to military sidearms.

The effectiveness of a ballistic weapon is proportional to the inertia of its payload (the bullet sans casing, the crossbow / ballista bolt, the rock in the case of a catapult). The effectiveness of an explosive is proportional to the mass of explosive times a constant based on the material used. Generally, well-controlled explosives (such as C4) cost more than poorly-controlled explosives (such as nitroglycerin); explosives with a higher yield cost more than those of a lower yield. Military organizations will opt for a mean between cost and effectiveness, varied based on politics and economic factors.

Is your society dependent on an effective military? If so, is this military large or small? If it depends on a small military, then there's a high expenditure per soldier, which means each soldier is well equipped; that means the weapons will be more effective. If it depends on a large military, then, ceteris paribus, each soldier will have less effective weaponry. In such a state, a neighboring power might provide weapons to those who do not mind flouting any bans on such imports--AK-47s rather than M16s in the US, for instance.

The relative security of military suppliers and borders will also determine what types of weapons are available. The existence of third-party suppliers that produce weapons specifically for civilian use is questionable; in a medieval society, you simply won't find people selling trebuchets to whoever wants one. Such engineers require patrons. But other weapons, such as swords, will be commonly made and sold. The general rule is, military sidearms are legal for civilians to make and own; everything else is highly restricted.

As for weapons and game balance, you're going to have to strain realism. A 9mm pistol is slightly more lethal than a 7.62mm submachine gun, but it's far easier to hit someone with the machine gun because you can spray bullets rather than aiming for each shot. So you could stipulate that the SMG does 12 damage while the pistol does 14 damage, but then nobody would use the pistol without a compelling reason. And you could provide that in the form of ammunition and the availability of the weapon:
- Only military personnel are likely to have SMGs.
- 7.62mm ammunition for SMGs is rare, except for military stockpiles. An individual soldier might have two clips when on guard duty, three or four on a day patrol, and perhaps a bit more for a longer patrol; a base might have five thousand rounds on hand, with more being dropped off every week or so.

However, since the 7.62mm ammo is actually larger than the 9mm ammo, you can artificially limit the amount the player carries, as well.

But as for the weapon designs, most will be entirely functional. Sidearms are much more decorative than main armaments, for the most part. The principle of pain is based on shoving material through a person to kill that person; that requires immense pressure, which is provided by giving a lot of force to a small surface area. You can use a tension force (bows), gravity (trebuchets), explosives (guns and, well, explosives), electricity (rail guns), or human muscles (blades and blunt instruments). Maybe you can exploit another force, though--be creative, browse physics books, and have fun! Then beware balance.

I think his point was that to maintain game balance, you shouldn't be able to always use the best weapons. Hence, artificial ammo scarcity.

I would agree that gameplay should be the primary concern when it comes to weapon design. Weapons have little influence on plot, or character, or any of the other "higher" elements of game design, but they are crucial when it comes to balancing the combat aspects of gameplay. They set the tone for fights and dictate styles of play.

If you fill the game with weak, rapid-fire weapons, players will wander around in a hellstorm of hostile fire, pausing periodically to line up a salvo of headshots or pepper an adversary with anemic lead. Colossally powerful, easily used weapons lead to paranoid creeping and a lot of mysterious deaths. Splash damage weapons keep players away from wall, while line-of-sight weapons have the opposite effect. More ammo means more fighting, and less ammo means more sneaking.

A "good weapon idea" is one that contributes to the game as a whole. Of course graphics and sounds are also terrific, but have relatively little to do with actual weapon design, from a gameplay stance.

In the world of the game I'm working on, there's a taboo firearms which keeps just about everyone within the realm melee combat.

So in response to Iron Chef, plot can influence the weapons to make the gameplay a lot more interesting.