At least I enjoyed myself until I didn’t anymore.

Hahahahahahaha! No. Tell the young’uns the truth, Sean.

When choosing a postdoc, to maximize chances for a future faculty position, you must, I repeat *must*, choose the institution with the greatest perceived prestige and which will give the maximum opportunity to get your names on the papers of famous, influential people. Then, when you get there, forget silly notions about establishing independence etc and devote all your efforts to attaching yourself to the research program(s) of bigshots, being junior co-author on their papers, getting them to like you and think of you as continuing their academic lineage.
If you don’t do that, then when applying for faculty positions you will join the queue behind those who did.

For a description of the postdoc life, with discussions of the challenges and strategies for advancement etc, I strongly recommend “My life as a quant: reflections on physics and finance” by Emanuel Derman. He discusses there how his fellow postdocs would cling to the elite institutions, even to the point of turning down paid postdoc positions at a lesser institution to work without salary at an elite one. Now why would they do that do you think?

Wanting to develop your own independent research program is admirable and all, but you mustn’t even think about that until you have tenure somewhere. Until then it is imperative to remain attached limpet-like to the research programs of the influential people who will be making or breaking you.

Let’s wind up Sean with a rhetorical question: In general, what will count for more when someone is being assessed for a faculty position:
(a) Single-author publications in PRL
(b) Junior co-authorships on routine papers of bigshots (along with lots of other co-authors)
(Hint: one of these will count for nothing; the other will count for a whole lot.)

As for my own experience: Only once was I in the fortunate position of having more than one postdoc offer to choose from. There were two:

(1) A 3-year offer from a very good youngish physicist with interests closely related to mine, who knew of and seemed to have a bit of respect for my work, was interested both in supporting it and interacting with me.

(2) An offer from a new assistant prof working on different stuff. He didn’t know or care much about my work, but was looking for someone willing to come and work for him on his project for one year, which was all he could afford to pay for from the money remaining in his startup grant. The hiree was supposed to take care of certain time-consuming but mundane aspects of the project, to free up time for the prof to focus on the interesting stuff. And if I wanted to continue my own stuff it would have to be as a hobby in my own time.

Sounds like an easy decision? You bet it was. Of course, I have omitted the crucial details: The first offer was at a lower tier institution while the second was at a semi-illustrious one. Needless to say, I chose the second one. The first guy was completely understanding, saying he understood that I needed to make the decision I made. And, looking back now, I can say with certainty that it was an absolutely correct decision. The relatively minor “stamp of approval” from having spent a short time at a semi-illustrious institution has helped me much more than any of the other stuff such as multiple single-author PRLs, plenary talks at major international conferences, and strong recommendation letters from leading figures in my acknowledged-as-”important”-but-less-fashionable-than-string-theory field. And in the quite likely event that I eventually have to leave academia it will count for even more.

I’ve also personally found that senior types that tend to be on hiring committees find these cross-pollinations particularly interesting.

Sean – in my experience it’s nearly always the number of papers that counts, with quality being of secondary interest. It angers me a bit, because it almost forces everyone else to focus more on quantity rather than quality. I guess it’s a consequence of the way research departments themselves get assessed.