The Trireme Veterans for Truth are back too! Hurrah!

I’ll have to remember to add them both back into the link panel. Their return is heralded by an analysis of Michael Shanks’s hit count. They argue from a very mathematically sound position that 409,908 hits divided by eleven months does not equal 60k hits a month, but rather 37,264. I think their analysis is flawed on two counts. One is that it assumes the hits are evenly distributed through the eleven months. If 37,264 is the mean point and the hits climb at a constant rate from zero when it started then it would be equally reasonable to assume archaeolog’s hit rate is 74,528 a month. 60k hits a month sounds plausible. Still, I can’t get too worked up about it because it’s irrelevant if you think about what a hit is. A hit is a file pulled off a web server. A web page like this one is purely text. There are no graphics in the file, so it counts as one hit. But you can see graphics can’t you? Like this one

There’s a bit of code in the web page that pulls the image file, in this case http://www.archaeoastronomy.co.uk/wp-images/smilies/icon_exclaim.gif into the page. Because that’s a separate file I’ve just doubled the hit count for this page. Add in the background, the bullets, the header at the top and once page can create a stack of hits. It’s hand if you want to keep track of costs for a web site, but it’s not really a useful guide to how many people are reading a page.

You could use the unique visitor count as a better tracker, but that still has faults. What is a unique visitor? You are. If you come back tomorrow you’ll be another unique visitor. In fact if you come back in an hour there’s a good chance you’ll be a unique visitor. I probably get a couple of thousand unique visitors a month, but it’s hard to tell. It’s a safe bet that of those visitors quite a few are the same person coming back every couple of days to see if I started to make any sense. If you exclude prospective vikings then I certainly wouldn’t think that I have a readership of thousands. I know Michael Shanks has a devoted graduate following, and there’s nothing wrong with that, but that might also inflate the readership figures slightly.

The biggest error that I see is that there’s an assumption made by both Michael Shanks and the Trireme Veterans for Truth that the visitors are human. That’s certainly not the case here and looking at Michael Shanks’s stats that’s not true for him either. If you look at the browser share you’ll see that 37% of browsers are from search engines. Visits from Google, Yahoo and Inktomi will trigger a count. Michael Shanks is not the only person not to realise this. Internet Archaeology has also counted search engine robots as visitors in the past. If you look at their logs then you’ll see the three most common browsers on their system are Googlebot, Msnbot and Slurp (the Inktomi bot). Around a third of the hits go on robots, which is a problem if you’re paying for advertising on the site.

But imagine we could make a sensible count of the page views would that matter? TVfT have linked back to my post announcing the return of Michael Shanks, but not everyone visiting their page will feel the need to click through to mine. Nevertheless the central message, that’s there’s something new to read coming to the net was picked up by the TVfT readers. Even if they’re not hitting my page, my message has hit more eyeballs. Equally you might not feel the urge to visit the TVfT post that I’m responding to, but their central message, that Michael Shanks’s readership claims aren’t accurate has also been passed on, even if you don’t read their webpage. I’d guess that means that both myself and the TVfT are entitled to add an arbitrary but ego-boosting amount on to our page counts. Page counts are not a useful metric, which is why sitemeter will disappear off the bootom of the page when I work out how to track referrals neatly.

A alternative could be to look at influence. Technorati ranks weblogs according to how many people link to a site. I get a rank of 147,261. The Trireme Veterans for Truth have a rank of 701,229. Michael Shanks is probably somewhere inbetween as he has two links. Even allowing for Viking wannabes I could still argue that with more links I’m more significant than the other two sites and thus the winner.

But it doesn’t work like that.

If I talk rubbish it doesn’t matter how many people link to me, I’m still producing rubbish. In fact if I wanted to kick up a fuss I could pull in some serious linkage and comments by being deliberately provocative. An example would be William Rubenstein’s post on Evolution at the Social Affairs Unit, which is massively linked to and commented on. Hardly anything linking to this is favourable though. They’re usually pointing out that even an Professor can be as ignorant as the next guy when stepping out of his field. Even linkage then is not a suitable measure of the importance of a site.

My definition of an important site is probably weaselly and shallow, but it’s the only one I know that can stand up to scrutiny. Does it have intelligent, well-thought out comments on it? Does it have a lasting effect on intellectual debate? Does it create discussion? There are a few sites out there that would pass this. Pharyngula is one, it’s one of the big spurs that started me on this because when I read some of PZ Myers stuff I wish I could write like that, and the only way to do that is practice.

I think Michael Shanks’s weblog is comparatively well-visited, certainly more so than mine, but not massively more. The Classics and Archaeology readership is rather independently minded and will find other sites if they have interesting material on them. If you don’t want the hassle of having a blog then Michael Shanks’s offer is generous and could be fun. But equally there’s no reason why you couldn’t set up your own site with minimal effort. Sharon Howard demonstrated this by setting up a blog in her lunch hour with Blogsome, which would give you a site a bit like this one. I think anything which increases the activity of intelligent debate about the past is a Good Thing, so it’s great to see two thought provoking sites returning to activity. —–