It is not very acurate. If your site in in the top 500 I am sure you are getting a lot of traffic. WW has a ton of Alexa users. More than most sites. That is why it is 419. I really don't think WW is neck and neck with USA today, the US Post Office, and the Drudge report. Even though that is what they show. I think WW broke into the top 200 back in Febuary. I know it is a busy site but it is still a niche site of webmasters.
Alexa is a decent indicator of a sites usage (at least top 50,000 type sites) if you consider whether or not the site would have an over average amount of users with the toolbar installed.
It is not a good idea to consider alexa rankings when buying a website because they can easily be manipulated. There was a lot of talk about people selling domains after having an army of offshore clickers boosting the sites ranking.
is a correlation chart put together by sites willing to submit traffic rank and actual traffic. Another data source that I sometimes use to correlate predicted traffic based on estimated web presence is data from the brochure of a banner advertising company with traffic for their client sites:
I installed the Alexa toolbar now since well 3 months. I'm working more than a full day on my page which shows Alexa rank 47,874 at the moment.
I can see some dropping when being away for 3 days to about rank 120'000. AWstats shows about 2000 different users a day and Nedstat shows about +4000 visits a day - not really changing when I'm away.
For well visited sites within the statistic of up to rank 100'000 I would say that there is a rather good indication about the traffic but not more. Up to rank 20'000 I would consider it reliable.
It is (today) not true that an administrator can run it down to 20'000 as has been said in other threads. We only can hope that this indicator for comparison of diffeerent pages can grow to a reliable tool for us.