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

> CONSIDER INVESTING IN RESPONSIVE COMPANIES

No.

As he notes, the sample size is ridiculously small. It's simply too likely that he got lucky with 1 or 2 companies in his "respond" group.

e.g., if he wrote these letters in the the early 90s and Buffet responded, he's immediately skewed his returns up because Buffet outperformed the market SO much in the 90s. That's likely luck, not replicable skill.

Otherwise, interesting article and the 1/3 response rate is surprisingly high.



maybe CEOs that are less negatively stressed by current affairs feel more free to answer him.. obviously this doesn't scale though


You can easily do a statistical signicance test to estimate if the effect is non-zero.


Which will still mean almost nothing, since past performance doesn't guarantee future results.


You basically just claimed that data analysis as a whole is invalid.


It's a valid point, this is post-hoc data analysis so you need to be careful of the Texas Sharpshooter Fallacy: https://en.wikipedia.org/wiki/Texas_sharpshooter_fallacy




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

Search: