[GMali]

And out of those, how many are parked?

[hessenwolf]

It is a study we would like to do. The plan is to select a random sample of domains for each length and query a source, e.g., Alexa, show a confidence interval by length for domain traffic.

The stats hypothesis would be to test whether length can be used as a predictor of traffic.

[janesvilleseo]

I was taken aback by the 13 character length, I wonder how many of those are parked, what the average length is for the top 100,1000, 10000 sites according to Alexa/Google/Etc. I suspect there there is a nice split between the parked longer names and the shorter actually used as a website names.

[nt_mark]

ycombinator - 11 facebook - 8 Wikipedia - 9 googleusercontent.com - 21

Shorter domains definitely make sense, but probably not a good predictor in itself though they have aesthetic value.

[nt_mark]

That's hard to analyse, however web traffic follows a power law distribution, http://www.hpl.hp.com/research/idl/papers/ranking/adamicglot.... So, outside the top 100k sites in terms of traffic you could expect it's either parked or idle domains.