Radiohead play coy on size of pot of gold at end of In Rainbows

Comscore's finger-in-the-air claims that about a third of people paid for In Rainbows seem to have become solidified as fact now, with the only debate whether getting more than one in three punters to pay cash for something they could get for free is a good result, or having nearly two-thirds of customers take your goods without so much as a spit in the spittoon is a terrible loss.

The figure is almost now gospel, despite Radiohead's attempts to deny it:

As the album could only be downloaded from the band’s website, it is impossible for outside organisations to have accurate figures on sales. The figures quoted by the company comScore Inc are wholly inaccurate and in no way reflect definitive market intelligence or, indeed, the true success of the project.

Which is in itself definitive, but surprisingly lacking on substance. Of course, Radiohead could blow Comscore out the water by releasing figures themselves, but they choose not to.

Which, of course, they're perfectly at liberty to do - although it hardly fits with the band's attempts to distance themselves from the pie charts and sliderules of the music industry.

Indeed, while Radiohead has been scrapping with Comscore, one of the big, old-style companies has turned up to hold the coats:

Warner Bros Records senior technology director Ethan Kaplan quickly savaged comScore’s methodology, which included only a “few hundred” in the study.

Oddly, Kaplan has never complained that the Grammys crown "best" artists and records based on a tiny sample, but never mind.

Comscore haven't taken this slight lying down. They've powered up their blog engine to launch a defence of their counting and stuff, and the idea of sampling in general. Who knew that Radiohead's album experiment would lead to statisticians rowing?