Hacker News new | past | comments | ask | show | jobs | submit login

I guess I don't entirely understand how this works - it sounds like you're introducing a param you don't actually have - you don't know what the clickthroughs are, how they're distributed, etc so this is entirely synthesized. Wouldn't you be able to simulate your way to more or less anything this way?

Seemingly worse for practical purposes, I don't think this data actually exists - adding clickthrough tracking to HN would be a huge change to the privacy profile of the site.




You are right. Our assumption for now was that clickthroughs are distributed like upvotes. But we don't have any data backing this up. That's why we need to do real experiments. And this touches to your second point that even on the real HN site, we don't have this data and don't want to track our users more than we do already. We're not entirely sure about it yet, but we could use an approximation instead. For example how long a story was shown on different ranks.

After all we got some very good feedback from HN and are incorporating it into our research. So thank you for it, this is valuable.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

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

Search: