I've been thinking a little about the combat mechanics of OD&D and AD&D and the way that those rules have developed over the years, particularly in the most recent editions of the game. Two of the pillars on which D&D combat rests are Armor Class and Hit Points. Both of these are abstractions, representing not a single "real world" thing but rather a concatenation of factors that, if separated and tracked separately, would slow combat to a crawl. Consider too that both OD&D and AD&D assume a one-minute combat round (as opposed to the 10-second one first introduced in the Holmes Basic Set and used in Moldvay, Mentzer, and 3e). One minute is a very long time in melee combat. During that timeframe, even the slowest, most encumbered person could likely launch a melee attack more than once. And yet OD&D and AD&D both allow only a single 1D20 attack roll during that time period.
D&D combat is thus extremely abstract; it's not intended to model individual thrusts and parries. Instead, it's the game mechanical equivalent of panels in a comic book, each attack roll being a distillation of many little offensive and defensive actions that occur over the span of a minute. The problem over the years is that referees and designers alike often forget these facts. Beginning with Supplement I in 1975, there's long been a tendency to want to "zoom in" on certain aspects of the melee round and treat them less abstractly. So you get rules for weapons vs. AC tables, hit location tables, combat maneuvers, feats, etc. -- all well-intentioned attempts to make D&D combat more "realistic," "fun," or "interesting."
Now, I don't think there's anything inherently wrong with the desire to get beyond the monotony of roll-to-hit-roll-for-damage combats. But the reality is that, if you're keeping Hit Points, Armor Class, and many other staples of D&D's original abstract combat system, many, if not most, "tactical" additions make no sense whatsoever. At the very least, these additions draw attention to the fundamentally "unrealistic" nature of the system, resulting in something that, in my opinion, is far less coherent. As I said already, this is not a blow-by-blow simulation of medieval melee combat, but most tweaks to the system treat it as if it were.
Given that, it seems to me that there are two possible approaches. The first is simply to embrace the abstraction and enjoy it for what it is. There's room for alterations and augmentations, of course, but they need to operate on the same level of abstraction or else the house of cards comes tumbling down. The second approach would be to revamp the combat system from the ground up, but that requires eliminating Hit Points, AC, and many other elements people associate with D&D combat. I think it could be done; whether result would be something anyone would recognize is another matter.
In short, I'm always amused when people speak of D&D as having a tactical combat system, because it doesn't. As originally conceived, the system was far too abstract for tactics to matter in the way most people use that term. But then I'm increasingly convinced that the strongest "philosophical" connection D&D carried over from its wargame roots was not in the realm of tactics but in strategy. It's this conviction that partly fuels my belief that OD&D was misread by a lot of the non-wargamers who picked up the game later. You can see the results of that misreading in late 2e, if not before, and it's in full flower in 3e (and almost certainly 4e as well).