Whenever this topic comes up, it is a bit abstract. I do think we have a problem with the "expertness" of Server Fault. That being said, in an attempt to start to pin this down I came up with the following expertness ranks for a question:

1. Person asking doesn't even know where to start, probably not an SA, likely to get a close vote or two but question stays open
2. Basic Question -- i.e. a junior SA or someone new to the particular technology. Answer is very likely elsewhere on the net or in a man page -- but that is okay
3. Normal SA question, understand the technology, but wondering how do non-obvious X in particular -- maybe to save time because someone has done it before. Possibly answered elsewhere, but not likely
4. Will likely require an expert in the technology to answer it
5. This is going to take a master in the technology to solve. Likely a bug, or a question particular a very small subset of users of that technology. i.e. high performance or issues with scalability.

Given a healthy site, and a healthy profession, I would expect to see a normal distribution of frequencies centered on Rank 3. This isn't based on any particular evidence, just what I personally feel the site and profession should be. Some might like it to be all 4s and 5s with maybe some 3s but this just results in comp.lang.c syndrome (Sick of "unprofessional" questionsSick of "unprofessional" questions)

Given what I know about system administration as field, I would still expect 3 to be the highest frequency, but with less weight on 4 and 5 and more on 1 and 2 if the site is performing ideally.

So what I did is going through the newest questions and rank them with quick snap judgments until I got bored. I plan on repeating this a couple times this week. This is, of course, pretty shabby sampling and highly subjective. On the other hand, it is at least some data.

Round 1:

Rank, QId
3, 291406
2, 291400
3, 291396
2, 291396
2, 291394
2, 291390
1, 291388
3, 291387
2, 291385
1, 291393
2, 291381
2, 291378
1, 291374
4, 291373
3, 291372
4, 291371
3, 291367
4, 291365
2, 291364
4, 291356
4, 291354
2, 291353
2, 291349
3, 291348
3, 291347
4, 291345
2, 291343
4, 291342
2, 291337
2, 291336
2, 291333