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

Was it? TBH I'm a little confused by the downvotes - going by the 100 sample size number @ a 95% confidence level this is a 10 point margin of error. This seems like basic statistics but maybe I'm missing something. I think it's fair to assume that they had people who were taught basic statistics doing these calculations. If these people then say they think the number of bots is 5% without mentioning the margin of error due to their methodology that seems deceptive. I'm not a lawyer but if I was on a jury I could see being persuaded by that.


> This seems like basic statistics but maybe I'm missing something.

Yep, the thing you're missing is that it was 100 samples per week, over a quarter, so 900 samples per calculation. I don't remember the margin of error cited but I think it was about .5%. Its not your fault though, because this is a misunderstanding that Musk actively encouraged in his tweets.

Musk has actually shown quite a few times that he either doesn't know what he's talking about here or is trying to mislead the public. For example, he also confuses total bot count and the number of bots in mDAU, which are different things.




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

Search: