You're actually doing a pretty good job of showing how well-balanced the weapon system is. The only problem is, you (oddly) seem to think "balanced" has to mean every weapon has to be about as good as another. The distincion between simple, martial, and exotic weapons allows for a bastard sword to be a better one-handed weapon than a longsword, and for a longsword to be better than a club. The warrior classes get access to better weapons than other classes, and if they are willing to spend a feat they get access to even better stuff. That's reasonable and balanced. The armor system is not.
Well, first things first. I'm not defending the idea that the armors are all equal. I'm defending the idea that there's nothing wrong with it (except the max dex on half-plate).
Now, to use the weapons table as our source, a quick look at it will reveal that there is no simple balance such that all simple weapons are equal and all martial weapons are equal, etc.
To list a few in the simple weapons category:
The light mace and sickle clearly superior to the punching dagger. The morning star is clearly the best one-handed weapon. (Although the heavy mace comes close (and may equal it if you consider metal-hafted to equal P/B), shortspear and club are far behind). Similarly, the javalin is better than the dart and the light crossbow better than the sling.
Among martial weapons, the light hammer is clearly the weak sister of the light weapons. The only thing to recommend it is its throwing range. Similarly, an unspiked shield is simply inferior to a spiked one.
Moving into one-handed weapons, the trident and flail are generally inferior to the longsword, battle axe, and warhammer. (They may be useful for corner cases like the warrior who wants to be able to throw his primary weapon or the warrior focussed on trips or disarms but those are not most warriors). In the two handed category, the greatclub, halberd, and heavy flail play second fiddle to the greatsword and greatclub.
Among the exotics, the gnome hooked hammer and dwarven Urgrosh are strictly inferior to the Orc Double Axe and Double Sword.
Admittedly, most of the suboptimal weapons (although not the greatclub) have unique advantages that may be interesting to some characters--trip/disarm bonusses or double damage when set against a charge for instance, but, IME, these are very rarely sufficient to make the weapons optimal. (Trip weapons are probably the only ones I think sometimes worthwhile).
The weapons may be balanced so that there is no clear winner but there are pretty clear losers in most situations. I don't see how this is that different from the armors.
As for your notion that a weapon that trades off a lower damage die type for a better crit threat range is somehow an inferior, "niche" weapon to comparable weapons of the same category, I don't see you providing any support for that assertion. It's certainly a subjective statement.
As for my contention that damage dice are better than crit range, the math backs me up. In terms of average damage per round, you need some very large plusses to damage before the higher crit ranger (or multiplier for that matter since it's mathematically similar) equals the higher damage dice. The only weapon likely to reach those numbers in the hands of many characters is a falchion in the hands of a power attack addict. Scimitars and Rapiers are very rarely equal to longswords, etc. And, even then, that only applies against creatures who are not immune to crits. And as fortification, constructs, and undead become more common at high levels, that makes a difference.
It [Breastplate] is hands-down superior [to chain mail], and 50 gp does not account for that difference.
Not for a 14 dex NPC fighter, paladin, or barbarian. Unless the character has a +3 or higher dex bonus, the only relevant differences between chain mail and a breastplate is one point of armor check and 50gp. For many characters the only one of those differences that matters is the 50gp.
The same is true of scale mail vis a vis a chain shirt. For a mounted fighter with a 16 dex, they end up with the same AC either way. The only differences are 50gp, the ability to sleep in the chain shirt (but lets face it, realistically, most people would not want to sleep in their armor anyway), manueverability (not very relevant if the character in question is primarily a mounted combatant), and one point of armor check penalty (which would be relevant when quickmounting or dismounting).
I'm still waiting for you to explain where folks have "missed the point" with regards to mechanical balance.
The point is that armors aren't "optimal" in a vacuum. They're optimal for a particular character in a particular situation. And many characters (like the 14 dex fighter deciding between a breastplate and chain mail) do not have the relevant abilities to make the difference between "optimal" and "suboptimal" significant.
The optimal armor for an 8 strength cleric is probably not a 50lb suit of fullplate. For a cleric who is staying out of melee combat, controlling weight may well be more important than maximizing armor class. And for a 10 strength cleric who does enter melee from time to time under the influence of various buffs, fullplate may still not be the optimal choice since its weight would be too much of an encumberance to allow him to carry a sword and shield without being heavily encumbered.
Armor class is not always the deciding factor when determining which armor type is best for a particular character at a particular time.
All you keep saying is that we're right. The chain shirt and full plate do outlcess the other armor types mechanically. Your only rebuttal seems to be "it's not all that much worse". The point is, they are worse and not in any significant way better.
I think you misunderstand what claim I'm rebutting. I'm rebutting the claim that the armor system is bad and that, under it, there is no reason for armor types other than chain shirts, fullplate (and, depending upon who's making the claim, leather, studded leather, or breastplates) to exist.
That the armor class afforded by the lesser armors is not always significantly worse than that afforded by the better armors is certainly relevant to the question of whether there's a reason for them to exist.
Same thing here. The fact that a character is apathetic about upgrading his or her inferior armor type does not really constitute any kind of counter-arguement, now does it?
That depends upon what you think it's countering. There are a lot of different semi-interconnected claims floating around in this thread and lumping all arguments against any of them together is bound to distort the argument. That characters might be apathetic about upgrading their defenses does nothing to counter the claim that there are clear optimal armors. It does, however, go a long way towards countering the claim that no PC or NPC above 3rd level has a reason to own/wear suboptimal armors.
And, I think the other implicit claim that I advanced--namely that optimizing a PC or NPC's equipment is not necessarily the same thing as optimizing their armor class is also relevant. The High Priest of Heironeous I described may well have optimized equipment for his position--despite having banded mail.
The positive claim that I and others have also advanced is also worth addressing. Since not all armor types have to be present in every region of every campaign world, the inferior armors may also serve a purpose
It seems to me that the grounds of the great armor complaint--the claim that armor types can only justify their existence if they are proven to give the optimal armor class for very broad classes of high level PCs in the circumstances of a standard dungeon crawl--is highly dubious. It ignores corner cases like the fighters and clerics I described, low level (poor) characters, NPCs with very limited resources, NPCs with limited access to high tech armors (like fullplate or even chain shirts in some cases), NPCs with equipment priorities not in line with dungeon crawling, and any cost/benefit analysis of animal barding (where cost is a far more significant factor and max dex is a far far less significant factor).