March stat pr0n.... not!

Look right, I don't do stat porn OK. I know I did it once with Alexa but I was bored and had nothing to post about. The reason I don't do stat porn is multi-fold really. Mainly, I am an old Bulletin Board/Usenet/Forum kind of guy. Stat porn is a bit like post counting. The more posts you do the bigger your name becomes in the "community" that you post and thus you get respect. This results in the "milestone post" where someone says "Hey I posted 1000 times! Go me!". It's a little bit like standing up in a pub and saying "I have emailed 1000 random chicks I don't actually know and will never meet on match.com. Go me! I'm dating!".

OK, so blogs are not quite the same, they're just websites, and with the growth of their usage, be it individuals or media, traffic is important. This is especially the case if you spend lots of money on it of course. You want to know that people are looking at the content after all. It's useful to know where your traffic comes from; it's useful to know how much traffic you're getting. In my job the latter is fundamental from a performance point of view. The problem is the assumptions that one can make about that data.

Now, the reason I raise this is because the irritant or as I like to call him the "Buy Canestan Cream!", of certain bloggers' existence - the ever so non-obsessive Tim Ireland - has posted a particularly good piece on the problem of stat porn. I'm not going to go into the detail, you can read it for yourself, but the key here is what his cheerleader Justin McKeating (no pom poms or short skirt included) has posted in the comments along with Tim's general point. You see, it's all about extrapolation.

I turned my servers AWstats on last month and in the last 14 days of March I got a total of 232,766 *hits*. Extrapolate that over 30 days and I'd have around 465,000.

Justin is of course absolutely right. Extrapolation, especially that which is linear, is universally stupid. It becomes assumption layered upon assumption, then wrapped within another assumption, at which point the idea that the answer is in anyway scientific, or dare I say it, truth becomes as risible as watching a man with only his right arm and leg taking part in a right-hand drive manual car race where everyone else has all their limbs. And no. I'm not ridiculing the disabled before you suggest it. Basically, it's called wit and you wouldn't understand.

Inevitable left wing non-obsessive criticism covered, I would like to take this opportunity to point out just slight - and possibly mild - inconsistencies in Justin's and Tim's posts about extrapolation. Whilst they're right to point out the very real problems with stat pr0n (along with the downwards, more serious, and statistically honest analysis of the figures), I'm still left feeling rather... shall we say... confused by their errant masturbation over their self-imposed genius.

You see, both of these "bloggers" (and I use quotes only because Tim does when describing anyone that does not fit into his view of what the definition is because he's not in anyway mad (or madder than me I should add)), are quite vigorous anti-Iraq war people. I cannot find reference on Justin's blog because his search facility is poo, but on Tim's he actively promoted back in August last year the "fact" that we must "feel the weight" of one million... yes 1 MILLION!... dead Iraqis and it was all out fault.

Where does that figure come from? Well... It comes from extrapolation that's where. It is based on nothing more that an extrapolated figure with another linear extrapolated figure rising upwards. You see it all began with The Lancet Iraq Body Count survey. The Lancet published a linear extrapolated survey of just over 1000 surveyed people asking them how many people they knew that had died a violent death in Iraq. Based on the response the extrapolation was made that approximately 600,000 people had died since the invasion of Iraq. Worth noting here that it was not an extrapolated figure because of the invasion, but an extrapolated figure since the invasion.

As one would imagine, the anti-war lobby leapt on this figure, and online they created a neat little time extrapolation that has increased the figure by piecing together random reports, records, and other 'evidence' to the point that the ticker rises as an almost constant. Put simply... it's bollocks. That doesn't mean that no one has died, it doesn't mean that lots of people haven't died either. What it means is that the figure is an extrapolated one based on a small sample, layered with further questionable evidence to suit a political purpose.

Now the latter point is absolutely fine in my eyes, the anti-war lobby is, after all, anti-war. They're fighting a political campaign and want to win, so I won't begrudge them doing their best with the data that they can find and manipulate to their political ends. This is politics right!? However, actively promoting extrapolated data whilst simultaneously complaining about extrapolated data elsewhere doesn't exactly place one in a strong position now does it? (I'm sure I've done it to on occasions I should add).

Don't get me wrong here guys though. I'm not saying you're wrong about the dangers and problems of stat pr0n. All I'm pointing out is that when you stand up and scream like big wet girls blouses about dangerous extrapolation and dishonesty - especially against those who are politically opposed to you - you should probably make sure that you've not committed the same sin yourself. It doesn't make you wrong on your point, but it does make you look rather silly, foolish, and some might even say big fat steaming great hypocrites (who are not very bight (maybe, I don't know?))?

Incidentally, before someone points out that because the Government has "accepted" the Lancet figure, that does not mean that it is right, or more correctly a truth. I'm not questioning whether one or 50 million have died in Iraq. I'm pointing out that the starting point for the figures is fundamentally flawed and so parading the figure as a matter of truth is bloody stupid. No. Actually. It's not stupid, it's retarded. I realise that calling Tim and Justin retarded might seem somewhat harsh, but whilst they might think they're on to a winner they're desperate quest for the "truth" misses two very core fundamentals, the first of which is hidden in Tim's own post about stat pr0n.

1. No one really gives a shit. OK? Seriously. We've got lots of people reading us all. They care, but most people don't, and when they see a post like yours - or for that matter a post like this one - they just scroll on by. Click, click, click.
2. More people will probably read this because they want to than will read yours. However, see (1) for how important that really is to the wider scheme of things.

P.S. Before it is suggested this is some sort of big defence of my blogging chums it's not (although obviously it will be for Tim because those are the rules of engagement and I understand that). The truth is, I just get overly irritated with people who can sometimes be quite clever being rather pointless tits (bit like me at the moment really without the preceding cleverness!). I also happen to have a soapbox upon which to shout about it (for no one to actually listen).

Here's a word of advice Bruce. Stick with issues you have about smears, and your (non-evidential) crap about sockpuppets. At least it had a modicum of intelligence to it (kind of) and you didn't have to contradict your own previous posts *wink*
Note: This is what is officially know on teh interweb as a 'flame'. I find Pinot Grigio helps.


Update: It would appear the battle of stat pr0n is now over at DK's. For the record, and if anyone is wondering, if you ask me face to face about my traffic I will give you a guesstimate. I'm rather pleased about it personally, for much the same reason as DK points out in the comments - sheer astonishment that it is as high as it is.