Revealed preferences are the preferences we demonstrate by our actions. These may be different from our stated preferences. Even if we’re being candid, we may not be self-aware.
One of the secrets to the success of Google’s PageRank algorithm is that it ranks based on revealed preferences: If someone links to a site, they’re implicitly endorsing it.
I got to thinking about revealed preferences when it comes to reference books the other day when I used some packing tape to keep the cover of my copy of Abramowitz and Stegun from falling off .
Instead of asking “What are some of your favorite books,” it might be more informative to ask “Which of your books show the most wear?”  This confounds frequent use and poor binding, but that’s life: there are always confounding effects.
My most worn math books are A&S, Bak and Newman, and Dunford and Schwartz. Bak and Newman was my undergraduate complex analysis book; I think it may have had a poor binding. Dunford and Schwartz got a lot of wear in college when I was into functional analysis.
I used A&S a lot in when I was developing a numerical library for Bayesian statistics. I still open it up occasionally, though not as often as I used to.
My volumes of TAOCP are in good shape, but I think that’s because they are well bound. I’ve cracked open Volume 2 quite a bit, though I hardly ever look at the other volumes.
What are some of your most worn books?
 Yes, I know it’s available online, but I prefer the dead tree edition. And yes, I know there are more extensive references, but in my experience anything I need that isn’t in A&S is unlikely to be in any other reference book.
 Benford’s law was discovered via revealed preferences. Simon Newcomb noticed that the early pages of a book of logarithms were much dirtier than the later pages. (Yes, Newcomb discovered Benford’s law, consistent with Stigler’s law of eponymy.)