I've never heard of anyone debating the World of Goo numbers before
It's probably because most people in the industry, naturally, would rather talk up the piracy figures than talk them down.
, and I'm not really sure what you mean by estimating the uncertainty- it was simply a statement by the developers based on the number of verified IPs that connected to their service versus the number that didn't.
It was clearly an estimate of the piracy rate. The headline was even simply "90%", put that way for effect.
It is common in the media for people to quote various point estimates of statistics without giving an estimate of the uncertainty. For example, many political polls with a two-party preferred question use about 1000 people as the sample and this usually results in about a plus or minus 3% as the 95% confidence interval, yet that little important detail is often omitted or at the best of times put in fine print on the television screen. So about 95% of the time, the true proportion of people voting one way or the other will be within 3 points of the actual claimed estimate. e.g. a poll that reports 48% prefer candidate A and 52% prefer candidate B would usually mean anywhere between 45 and 51% prefer candidate A, and 5% of the time that would be wrong.
In the article about the piracy of world of goo, no attempt is made to estimate the uncertainty in the estimate. To me, this is usually one of the biggest indications that the author is not qualified or at least not justified in making the statistical judgements they do.
He does make a point to call it a "rough estimate", but the damage is done and people will then gobble up that number and use it in debates like it's a scientifically proven to a high level of confidence "fact".
It's been a while since I actually looked at it, but nothing about it seemed manufactured to me.
It doesn't need to be manufactured for it to be wrong. When you look for statistics that suit your argument, it's very easy to find them even if the argument you support is probably wrong. This is why in medicine there are such stringent conditions for how to use evidence from experiments. It's called Evidence Based Medicine and IMO it's one of the reasons Western medicine is such a success.
Yes, the level of evidence in examining piracy rates may not be quite as important because it's not as if lives are depending on it, but there still should be minimum standards for the accuracy of evidence IMO.
As for the indie games, it's been my experience with titles of this nature that people typically don't bother pirating them. They usually appeal to a small fanbase of more dedicated users who actually support the makers of the game. Take Dwarf Fortress, for example- I can't imagine too many people play that game considering it's complexity and the fact that the graphics are all ASCII art, yet the developer still makes a living solely off donations. This is all conjecture and personal experience though, there really isn't much information at all about piracy rates unfortunately.
Agreed. All speculation, and pretty much always anecdotal evidence dominates these discussions.
.
