A complaint I have heard before from new players, particularly martial artists, fencers, and SCA types, is that it's too damn hard to hit in D&D (especially at low levels). You only get one attack every six seconds, and you have a 50% chance to hit an unarmored person? What the hell? Watch this! *bap bap bap bap bap bap bap* The crowd that plays D&D to emulate myth and media runs much the same objection - heroes in their source material are deadly-accurate.
But I think it's actually pretty reasonable. Like the Original Megadungeon, I think military culture and fighting experience informed the design decision to have most combatants in D&D be quite incompetent.
Consider the Marshall Study. SLA Marshall in World War II documented that 75% of US servicemen were unwilling or unable to fire to kill on the enemy in combat. While Marshall's work is controversial, it led to changes in training for American infantry that boosted their "attacking percentage" up closer to 90% in Vietnam. This is, coincidentally, the correct timeframe for the Ur-Gamers to have heard about Marshall's research, and to have modeled the poor THAC0 progression off of it.
Another source which supports Marshall's claims is Collins' Violence: A Microsociological Theory. This is a very long and exceedingly bleak book, and I have not gotten very far into it, but Collins' method is the analysis of footage of fights, played in very slow motion, so that fighting can be observed in detail as it truly is, rather than as it is recounted. It turns out that people lie about fighting bravely, and Collins' analyses of combat footage largely agree with Marshall - there is a particularly damning still, where a group of armed men is under fire, and seven seek cover behind each other while only one returns fire. Most people shy away from inflicting effective violence when they can see the face of their target, unless they have a bunch of other people backing them up and cheering them on, and this is true across a number of observed cultures (not just modern westerners). Most effective violence is inflicted with four-against-one or greater disparities of force, or against fleeing foes. This is, incidentally, why the pursuit of a routed army was where most combat casualties occurred in Classical campaigns.
So yeah - your average 1st level henchmen would be lucky to make a credible attack every six seconds. He's going to spend a lot of time hemming and hawing and evading, and his opponents are doing likewise, because neither wants to kill or to die, and neither has had the sort of conditioning employed by modern militaries to make it easier. Psychologically, fighting hand-to-hand with real weapons for life-or-death stakes against a hostile opponent is incomparable with fencing in a controlled environment, and it introduces a whole new layer of stress and adrenaline and clumsiness (another of Collins' findings - people fall over, hit the wrong targets, and generally fumble and flail an awful lot in real fights. It has me reconsidering my stance on fumbles on natural 1s on attack throws). That is the nature of the beast called man.