Probably the biggest single source of bad rule and mechanics decisions comes from the fact that most game designers, rather than actually going to a library, base most of their research on the work of other game designers.  In this way errors are compounded, unrealistic ideas are perpetuated, and design flaws from the earliest of games become commonplace in all modern iterations.  It all goes back to the origin of the "role playing game."  Here we are talking about the true origin of the "let’s pretend" game like House or Cops and Robbers, but of the origin of the systemized, rule-based role playing simulation.  It all starts with Chainmail.

Chainmail was a short, cheaply published book by Gary Gygax and Dave Perren, originally published by Guidon Games, copyright 1971 (Gygax claimed 1969, but the copyright information contradicts this).  It was a set of rules for tabletop miniatures battles using lead figures and dice, and contrary to popular geek-convention myth, there was indeed a 12 page fantasy rules supplement present (in addition to the now-standard concept of "hit points").  By 1974-1975 it was being published by Lake Geneva Tactical Studies Rules (TSR).  About this time, the first version of Dungeons and Dragons came out as sort of an add-on to Chainmail, three little brown books that focused on the playing of individual figurine-characters as opposed to conducting large scale tactical combat.  The combat tables determining hits, misses, and damage were very similar to those in Chainmail. This is where all the trouble begins.

The modern idea of the systemized RPG, from pen and paper to MMORPG, all stems from this Chainmail legacy, and several silly factors have never been properly weeded out.  The two biggies are:

• The focus on combat as the core activity in games that are purportedly about assuming a role, as opposed to a tabletop miniatures battle system
• Arbitrary and unrealistic characteristics ascribed to weapons and armor, very convenient for calculating casualties in mass combat quickly yet wholly inappropriate for small-scale tactical simulations

It is difficult to underestimate the ability of people to consider themselves an authority on things when their only source of information is a game manual someone else wrote before them, which was in turn based on another game manual, etc. etc.  Things become worse when these designer-types elect to "make a few little adjustments" in the thirdhand systems they’re stealing from.

Example:  In Advanced Dungeons and Dragons, 1st Edition, there is a huge hairball of a combat chart that nobody used, called the "Weapon vs. Armor Class Type" table.  This one table helps to make sense of the arbitrary D&D weapon damage system of, "duh, a dagger is pretty small so it does 1d4, a bastard sword is a little bigger than a long sword so it does 2d4 instead of 1d8…".  The table assigns various bonuses or penalties to hit with any given weapon vs. an armor type, i.e. a weapon may have terrible chances against full plate and shield, but be better at penetrating chain.  The table was flawed (it treated chain and shield the same as it would treat splint, as they were both base AC 4, etc.), but it lent some purpose to a weapon one might not otherwise consider useful.  It was also very complicated to use for the typical beer and pretzels gamer, and so nobody ever used it.  In the misbegotten later editions of AD&D, this table was simply removed altogether, and players were back to the old choices of longsword, longbow, and two-handed sword.

The advent of the computer as a referee, or even a referee’s assistant, presents entirely new possibilities to get away from these godawful abstrations and arbitrary damage values.  A computer is perfectly happy to calculate whether or not the sun is shining at a bad angle into an archer’s eyes, or the effects of windage on a sniper’s shot.  As better, more powerful machines become available and come down in price, the potential of the MMORPG server to do complex battle calculations increases.  The basic ones aren’t even that complex.  Is the guy being stabbed by the stiletto wearing maille?  Then the stiletto has more effect!  Is a spear longer than a club?  Advantage spear!  Simple considerations like these can add a whole new level of subtle realism to the game, and with it, a whole new importance to strategy in combat, as opposed to how high one’s numbers are.  If game software engineers would stop thinking in terms of an expensive version of 3 little brown books published in 1974 and more in terms of computer-simulated battle conditions, I would be ecstatic.