A word on "methodology", i.e. wot I done.

From the source data (Excel file AnlaegProdTilNettet.xls alleged to come from the Danish energy web site), I filtered the offshore projects. The data is presented as production figures from individual turbines, but in reality it's clearly some sort of composite figure, as each machine of a given type in a given park has the same numbers. So I chose one line from each group. I've double checked, and in the source data, the projects aren't referenced by their habitual names, but the correspondances are clear with a bit of digging.

In the above graph :

• Uoplyst = Vindeby
  
• Ukendt = Tunø Knob
  
• 2nd and 3rd Uoplyst = Middelgrunden
  
• Horns Rev = Horns Rev
  
• 2nd and 3rd Ukendt = Rønland 1
  
• Hav = Samso
  
• The last Uoplyst = Nysted

So this graph contains no aggregation or averaging of any kind, other than that of the source data, and represents capacity factor by year of all the pre-2009 Danish offshore parks.

I eat raw data for breakfast. Great bleeding chunks of it. But I'm a moralist, and I don't like to see data tortured. Clearly the above data must have suffered a great deal to get it to "confess" the trend that Hughes gets from it. Because the trend just isn't there.

There are too many British onshore projects to submit them to the same naive treatment, but from looking at the data, the story is the same.

It would be worth keeping an eye open for press take-up of this atrocity, in order to "set them straight". Better would be to pre-empt them of course...

A quick google shows that there hasn't been much take-up yet, though the FT has published an article, which I am composing a reply to.

_{It is rightly acknowledged that people of faith have no monopoly of virtue - Queen Elizabeth II}