One of the many things I've learned to embrace in preparing my Old School campaign has been what Mr. Maliszewski over at Grognardia refers to as "The Oracular Power of Dice". I also got the name for my blog from a post by Ben Robbins at Ars Ludi on a similar topic. Basically, the idea is to accept what the dice give you and then craft your plans around them, rather than to have a master plan and only accept results that fit with that plan.
Now, this is a little bit hard for me because I tend to have some pretty definite ideas about the "flavor" I"m going for, and I tend to want to make decisions about everything. But I've discovered a few interesting things while prepping Fellhold. First, it is strangely liberating to just go through rooms and roll to see if anything is in there and what it is. It removes a lot of the burden of trying to think up logical monster and treasure placement and hold the entire dungeon ecology in your brain. It also removes worries of being either too easy or too harsh on the players, as much of it is out of my hands.
So, ease of use and making the ref's job easier are widely known and pretty obvious benefits of random monster and treasure stocking. What is perhaps less obvious is the way that taking this unexpected input and linking it to what is already established results in unexpected depth. Sometimes asking yourself "okay, so why are these hobgoblins in this room?" with them being there as a given inspires more creativity than thinking "what is this room and should it have hobgoblins in it?". Also, sometimes strings of monsters will just seem to fit so perfectly that you can't believe you didn't think of it on your own.
Finally, the non-random aspects of this process are interesting as well. I'm learning to love stocking tables and wandering monster tables, and figuring out cool things I can do with them because of it. So, originally I was just using my wandering monster table to stock a particular sublevel, but for further levels I'm thinking that I will have separate room stocking tables and wandering monster tables, since different sorts of monsters are more likely to be wandering around than others. Also, it helps me to establish the sort of ecology that exists in a particular place without actually having to establish every little detail. I'm planning on playing around with multiple dice tables, so that I can have more nuance in frequency (for example, use 2d6 so that things in the 6-9 range are more common, whereas the 2 and 12 results are quite rare). Hopefully if I can develop sophisticated enough tables and charts, I'll get to the point where all I have to do is draw a floorplan, plop down a few points of interest, and then roll up a totally fleshed out dungeon level. Such techniques will also help when we get to wilderness adventures.