If some of your games are under 50k, I guess +/- 10% is reasonable.Ĭonsidering that the data doesn't belong to the publishers or developers, I see absolutely 0 reason for him to comply if he doesn't want to, the information is not gathered from anything that the publishers/devs have rights to, at best valve could complain about it because it is mined from the users of their service. For very low sales (say <10k), Steamspy is widely innaccurate outside of a "That game didn't sell" perspective. It's the same methodology, but with around 20k instances to worth with, private profiles and other factors will affect the derived user total more than it would a million seller. That's why something like Trails in the Sky SC is sitting at 22,458 ± 3,819 (+/- 17%) on Steamspy, while Undertale is sitting at 1,859,055 ± 34,655 (+/- 1.8%). With sampled data, larger samples tend to be more accurate than small samples.
So, 10% is 10% regardless of what it is 10% off. It doesn't matter that I was off by $250M in the first case and only $30k in the second, because that $30k is much more important to a $100k total than a $10B total. If you were a billionaire worth $10B, and I estimated your net worth at $9.75B, I would be giving a much more accurate estimate than I would be a case where I estimated your $100k net worth at $70k. Accuracy is just a percentage deviation from the actual number.
0 kommentar(er)
