On the contrary, I see the "three little booklets" of Original Dungeons & Dragons presenting a very different picture. I see that old White Box as a complete game, much like boardgames contain a whole game in one box. There is a beginning, a middle and an end, and you are expected to play from start to finish almost every time. The secret to why this never seems to happen in later games, I believe, is twofold: there is the problem of statistical lethality and the problem of pacing.
Let me be explicit: the dungeon delve as exceedingly dangerous to the heroes is part and parcel of the old school experience. Not only is the dungeon itself "inimical to men", but the quest for treasure all but guarantees flashing swords with treacherous and unspeakable monsters that are far more powerful than the heroes. Indeed, the sudden reversal of fate or noble sacrifice, the loss of a brave hero, is an essential part of the story.
But there is a difference between a sudden upset and statistically impossible survival: there is nothing heroic about a bunch of low level characters in leather armour getting killed. It's neither surprising, interesting nor the fuel for a good story.
I believe OD&D has one rule that is quite often overlooked. It is true, the White Box introduced the "d20" attack roll which has become the flagship of every edition of the game since, but in reality this was merely a provisional rule, a stopgap measure since the game actually required the Chainmail rules. It is quite possible, in fact, that the d20 attack chart was quickly thrown together to smooth over the fact that the original game was shipping as an incomplete game, requiring expansions right from the start.
I will give a more complete treatment of the venerable tabletop wargame Chainmail in the future, but it's worth looking at the Man-To-Man melee and missile rules that were intended for OD&D. A foe with a spear has more than seven times higher chance to hit a fully armoured combatant with the d20 system than with the Chainmail system (where he only strikes on a roll of 12 with two six-sided dice). The common shortbow produces identical numbers. With plate armour and a shield being easily affordable, a Veteran (1st level Fighting-Man) can expect to survive an average 72 attacks before succumbing to a fatal wound. This is not to say a lucky shot cannot end his career much earlier, of course, but the thrill of the battle should be that the arrow to the heart is possible, and thus tragic and momentous, but not taken for granted.
Additionally, while the three little booklets make many references to Morale checks, Original Dungeons & Dragons leaves enemy morale rules to Chainmail. It is worth noting that, with these rules, many low-level foes will flee after taking a small amount of casualties (usually one third losses). This is important because only in later editions of the game did a concept of "killing for experience" develop, where you only earned reward for foes that were slain. As a wargame, total victory in Chainmail is most often marked by routing your enemy, meaning that this should also be considered a complete victory in OD&D and thus worth the total experience reward for every foe so vanquished.
What particularly stands out with the original game's experience system is that experience gained from combat is about ten times faster than AD&D at earlier levels and actually slower at later levels. With the average 1 hit die foe worth 100xp (and even fleeing foes worth full experience, as suggested above), the heroes will attain second level after only a few harrowing battles. Much more significant discussions of the wisdom behind this system have been levied before, of course, so I want to focus on experienced earned from treasure.
Like many early editions of the game, Original Dungeons & Dragons awards experience for fighting off foes and unearthing treasure. While there is no explicit guidelines for treasure, referees will often settle on the AD&D standard of 1xp per coin (although Gary himself considered 5xp per gold piece in his last manuscript, Castle Zagyg: The Upper Works). Unlike later editions of the game, fighting combat and digging up treasure are always separate events in OD&D. Monsters do not have pocket change (later editions would give many humanoids "lunch money"), but rather hide their fortunes away in lairs (I would go so far as to say that money found otherwise is not worth experience). This is important because negotiations are more weaved into the fabric of OD&D than in any later edition, meaning treasure (and experience) was potentially much easier to obtain. Not only do the rules imply that most monsters can be bribed, but the game actually assumes the heroes will sooner or later have a full retinue of monsters and mercenaries at their command.
The biggest argument for faster paced treasure experience is the implicit endgame in Original Dungeons & Dragons. Wherever you turn, you come across the assumption that heroes will gain fame and fortune and build a fortress, which can easily cost around 100,000 gold coins. Often, referees are far too stingy with treasure, but OD&D has a very different economy than later games. If a hero is to have enough money to construct a castle by level 9, and each coin earns a point of experience, then easily half the experience for each level should be in treasure. This means for a Veteran to become a Warrior, he should ideally have personally defeated 10 Orcs and retrieved 1,000 gold coins.
This conclusion is merely drawn from the game implicit in those original little brown booklets. How often have you actually played any edition of old-school D&D where you started with the best armour, had a 3% chance of suffering harm, were able to withstand two wounds on average, fought off 10 Orcs, successfully bribed your way past other foes and exhumed 1,000 gold pieces for your effort? If you answered "not often," this is most likely a result of our forgetfulness of the implicit endgame in Dungeon & Dragons. I always argue that the earliest examples of roleplaying games are more boardgames than anything (freewheeling and off the cuff, to be sure): they are meant to be played through to the end. It is only the incidental inclusion of a polyhedral combat system that turned heroic exploits into essentially a medieval horror game (a legitimate conceit in its own right, to which we owe much of the compatibility with Cthulhu themes). When tied to its wargaming roots, we uncover a play style that is much more evocative of early Greyhawk, with its knights in shining armour, damsels in distress and fiery dragons (which all but disappear from gaming tables with the death of the endgame).