Originally posted by beholder3 At least the massive collapse of Nikon sales reported by BCN in Japan:.
Look, I usually enjoy reading your provocative and informative posts, Beholder3, but since I've already taken Mistral75 (who is an honourable gentleman) to task over this same methodology, to be fair, it's your turn.
This famous news photo says it all about what happens when the proportions in your sample fail to represent the proportions in the population as a whole:
The very expensive phone polls didn't cater for the fact that not all the voters on that Tuesday in November could afford telephones - the less affluent Democratic Party supporters were less likely to be surveyed.
Like the BCN stats you rely on, it's a great example of 'selection bias' ...
Selection bias - Wikipedia
JPT is modest in saying he doesn't know much about BCN, but clearly he does:
"BCN Ranking lists 23 retailers on its website. The ones of note for cameras are Bic Camera, Amazon (not as big as in other countries), a couple of more specialized camera retailers, a number of electronics chains that are not so serious about cameras, and a few other online retailers. I think that they have around 50%, but more towards the consumer end. If they claim 60%, it is probably a figure for the electronics market as a whole. On the whole, these are just not the shops people go to by a Pentax, or any specialised camera."
I know you've gone to a lot of effort, multiplying BCN percentages by CIPA figures and graphing them (you've put them on several threads), but they're *invalid*.
I'm assuming you have a scientific or engineering background, as I do.
What would your professor say of the lack of error bars - the uncertainty in your figures - when you confidently declare Nikon sold 248057 units last year?