Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Hard to cheat statistics.

This is how they find accounting fraud as well.



No - this is how they find suspects for accounting fraud. They still need to show actual proof of fraud.


How you suggest they do that in online tours?


It's actually fairly easy to cheat statistics. It happens literally all the time in Academia. There's a thousand ways to make a statistical analysis believably say what you want in a way where even other professionals don't realize is the case unless an expert does a thorough analysis.


It's easy to cheat a statistic you create. It's quite hard to cheat a statistic where you don't know who will look when at which particular data points.


Which is exactly the case here, as they decided what data to look at and how, and possibly had a bias.

Further, clearly the analysis wasn't so irrefutable given that they admitted faults with it after others pointed out mistakes[0]

0. https://twitter.com/IglesiasYosha/status/1574308784566067201?


I did not intend to attack or defend Hans with that statement, I just wanted to point out that both you and the original comment could be right at the same time. That being said, it's quite funny that this case showed both sides.


Yeah, but it is also extremely easy to misrepresent statistics!




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: