I have a hard time disagreeing with that, since one of my guiding principles, higher even than fidelity to pulp fantasy, is that D&D is always right. By that I mean that, before considering changing, let alone removing, some element of the game that's been around for decades, look carefully at why this element was introduced in the first place and consider just as carefully what a change might do to the gameplay of D&D. Basically, "D&D is always right" is a reminder that, though the game may look haphazardly constructed, there is a logic behind it, a logic that you really have to understand and respect before you can start monkeying around with its internal organs.
As he so often does, Philotomy sums up the case against the thief as follows:
The Thief class is not part of the original three OD&D books, but was added in Supplement I. Weak in combat and casting no spells, the main feature of the class is its special skills like climbing sheer walls, finding and disarming traps, moving without making a sound, hiding in shadows, executing surprise backstabs, et cetera. Over time, I've come to prefer the game without the Thief class (i.e. using only the original three classes). The role the thief usually plays (scout/sneaky-guy) is easily filled by the other classes; everyone can attempt to be stealthy, search for traps, et cetera. Also, without the Thief and his special abilities, these activities are often performed by the player describing how he goes about it, rather than rolling against a skill, which I think is a lot of fun.I'm largely in agreement with Philotomy's assessments, but I've nevertheless been reading Greyhawk's description of the thief very carefully, as well as the Holmes rules, looking for insights that might help me "rebuild" the thief into something that's simultaneously true to the archetype that many now (rightly) consider a staple of D&D while still taking into account the (valid) criticisms many cite against the thief.
I have noticed a couple of interesting things. First, the OD&D thief, unlike his AD&D descendant, has no ability to find traps. Or rather, he is no better at finding traps than any other character class, who use the standard "rules" for doing so (i.e. player deduction). Now, I did know that this was the case and it makes sense, given that several people, Gary Gygax chief among them, have noted that the thief owes its origin to the desire for a "dungeon bomb specialist" class, who was better at removing traps than other characters. Second, in OD&D -- again, AD&D differs in this regard -- a thief's ability to hear noise uses the same game mechanics as the standard rules describe in The Underworld & Wilderness Adventures: a D6 roll. Holmes further explains that "The thief's ability to hear noise at closed doors, secret panels, etc. is rolled on a six-sided die like anyone else, but his ability improves as he advances in experience." (emphasis mine) Indeed, until a thief reaches third level, he is no better at hearing noise than is any other character.
Taken together, these two things might point the way toward making the thief more palatable to me. Some have suggested in the past, Philotomy among them, that it might be best to view a thief's abilities as extraordinary in nature -- being able to climb sheer surfaces rather than ordinary walls, for example (which is exactly how Moldvay does it). Taken this way, the thief's abilities are more akin to a magic-user's spellcasting; they're "powers." While there's nothing wrong with that approach, it doesn't appeal to me, both because I don't think the mechanics for the class, as presented in Supplement I, support it and also because it's the first step on the path to turning the thief into the ridiculous ninja he becomes in the WotC editions.
To rehabilitate the thief in a way that's in keeping with its origins, both in Greyhawk and in pulp fantasy, I would much prefer to treat the abilities of the class more like the combat abilities of the fighting men -- ordinary abilities at which the class excels compared to other classes. The difficulty, though, is finding a way to do that without fostering the inevitable diminishment of the ability of other classes to attempt -- and succeed at -- those same tasks. I'm not yet certain that it's possible to do that, but I am exploring that avenue and will discuss it at greater length in future posts on this topic over the coming days.