The WSJ (which requires your first born to view an article), summarized some work by Dennis Forbes.
Some interesting points from the article:
The most common word four letters or longer, though, is “home”; 719,000 domains have some sort of home in them. Given the economics of the Web, chances are that many of those involve refinancing: 114,700 URLs mention “mortgage,” which is more than discuss “science,” “nature” or “children.”
Half of all domains are between nine and 15 characters long; the average length is 13. A domain can have, at most, 63 characters, and there are 550 such domains. In fact, some people have made a haikulike art out of 63-character domain names.
While much has been made of domain names like business.com being bought and sold for millions of dollars, Mr. Forbes is dubious about the value of expensive domains. Most people now search for Web sites using a descriptive word or phrase, or else are introduced to a site by a friend or colleague who emails the URL. So, domains don’t need to be short and snappy the way they had to be in the earliest days of the Web.
This is from Dennis’ original article
Stepping up to four letter sequences, choosing among the 456,976 combinations, yields a vastly greater availability — perhaps the set is a bit too large for domain speculators and their unlikely success with random sequences — with 97,786 showing as open. A quick check verifies that most are legitimately available. “Choice” domains, such as AGJV.com, EIYK.com, GZVW.com, and QFEV.com. Adding digits into the mix and there are a massive 1.16 million open domains, so long as you’re looking for something like 7RG8.com, or U3JZ.com. Choose one and then manufacture a ridiculous backronym to explain it.
There’s a lot more in the article worth reading. Definitely worth a visit.