Lies, d****d lies, and statistics

Much is made on the news/rumor sites (and almost weekly by one Youtuber) of which camera manufacturer is ‘top’ in sales.

The monthly CIPA stats, for instance, tell us the numbers of types of camera produced and shipped. (Note ‘shipped’. That’s not the same as ‘sold to customers’ but they are too often taken to be the same. Obviously some cameras probably leave a retailer very soon after they’ve arrived—we can all think of which ones—but others may be added to stock.) But they are not broken down by manufacturer.

The other stats often relied on are usually from Japan, either the cumulative numbers of sales from a group of Japanese stores or just one. These vary markedly from month to month (or in the case of Yodabashi, fortnightly) and tell us nothing more than what happens to be either popular during that period, perhaps most available, or even possibly most affordable thanks to cashback or sale offers or price reductions.

The other problem, of course, is quoting rankings where X is first out of ten, and P is ‘doing badly’ because it's tenth. Retailers are not going to (and never will) tell us that X is top because a hundred examples were sold and P is bottom because its sales amounted to 84. Or X was eleven and P one.

And of course, those rankings and figures apply only to Japan. Where, unusually, fixed lens digital cameras far outsell mirrorless ILC for example and DSLRs appear to be very near extinct.

The information on what particular brand (let alone camera or lens) sells best, or is most popular, can’t be derived from any of these stats unless you at least look at the numbers or rankings for an entire year.

They may fuel ‘fan’ enthusiasm, or contribute to soothing the ‘Was I right to buy X’ anxiety syndrome, for a day or two but the oft-quoted stats are mostly pointless.

The most egregious error was perpetrated recently when more than one site claimed the CIPA stats showed mirrorless camera sales (see above, actually ‘shipments’) had increased by ‘112%’ over the previous year. No, they had not more than doubled: the increase was 12%, since the previous year was used as a baseline of 100.

Am I the only one to be irritated by this misuse (or misunderstanding) of statistics?