Statistics are great to explain things, to show trends. When the numbers are understood, the statistics based on those numbers provide excellent value.
Nikerabbit is working on the Betawiki statistics at the moment. I do not know all the goodies he is planning, but the first batch is really interesting. Niklas has added statistics to the portals for the languages. When you look at the Dutch portal for instance, you will find under the Babel values a graph with the edits pers day over a period of a month. I think Niklas put it in a reasonable place, it gives a clue as to the recent activity for a language, and I can imagine that it helps to keep people interested.
I have also been given a static graph of the number of edits. This shows how the activity on Betawiki develops. When you compare these numbers with the well known group statistics and the group statistics in time, you get the feeling that when a language is more or less finished, the involvement in Betawiki wanes.. What you can observe is that the support for a language often happens in spurts.
Now that the Wikimedia Foundation has hired Erik Zachte to work more of his statistical magic, it will be interested to learn if he can find a relation between the localisation of MediaWiki and the number of editors, the number of readers of a project. It seems obvious, there is a lot of circumstantial evidence, but it will be really important when numbers support what is to me a conviction, a conviction that is shared by all of us here at Betawiki.