In the nearly two years I've been writing this blog, a lot of (virtual) ink has been spilled about the thief class and its suitability in old school D&D. On a largely theoretical level, I've been more or less firmly committed to the "Just Say No to Thieves" camp, seeing the class as an unnecessary interloper whose presence lays the groundwork for a number of later unhappy developments. But theory that doesn't work in practice is bad theory and, as I've discovered over the course of the Dwimmermount campaign, the notion that the thief class is a priori bad is mistaken, or at least it is often mistaken.
Brother Candor employs a cynical, pipe-smoking woman named Gaztea, whom he hired in Adamas primarily as an information gatherer. However, she's also technically a Supplement I-style thief, which I allowed as an experiment to see whether her presence would affect the game in any significant way. Well, Gaztea has affected the game but not in any of the ways I worried she might. For one, low-level thief abilities are pretty hit or miss, with the exception of climbing walls. Second, Brother Candor and the other characters rely on Gaztea far more for her social skills, which is to say, her underworld connections in Adamas, than they do for her skills in picking locks or finding traps (She's also, of late, taken up the study of ancient Thulian and alchemy, but that's another story).
In many months of play, none of the things I feared about thieves ever came to pass. A big part of that, I think, is because I'm aware of the potential pitfalls and have done a good job of avoiding them. Equally important is that my players, especially Brother Candor's player, keep Gaztea's abilities in the proper perspective. They have the right frame of mind and don't see thief skills as an excuse for non-thieves not to try "thiefly" things. Indeed, I'd venture to say that Brother Candor, Dordagdonar, and Vladimir all spend more time finding and removing traps, for example, than Gaztea ever does.
The example of Gaztea is a good reminder to me of that eternal truth: the play's the thing. I can muster 1001 cogent arguments against the thief class from a theoretical perspective, but none of them are worth anything if, in actual play, the thief class manages to avoid all the dire consequences I "knew" would transpire. My experience also serves as a reminder on the design side of things as well: don't design rules with bad players in mind. Any rule, no matter how well written, can and will be abused by players looking to find a way to take advantage of a situation. I think many (though not all, by many means) of the arguments against the thief class arise from a fear of abuse by players or even referees and, while one should be aware of such things, I think it a grave mistake to make decisions based largely on how a rule could be abused.
In the Dwimmermount campaign, despite my skepticism, the thief class is working, so I'm going to continue to allow it, not just for NPCs but for any future PCs as well. Obviously, every referee must make a determination for his own campaign based on his own experiences, so I don't mean to suggest my stance ought to be universally adopted. Still, I have to admit to some small joy in discovering my long-held skepticism was misplaced in this particular case.