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

I don't think his analogy to reputation economies is quite right: having been a Lyft driver, I did observe that it was difficult to get out of ruts when you have a low score (a passenger entering a car with a low score is going to be more vigilant and critical). Yet, scores are constrained to being between 1 and 5 so it's relatively easy, with the right strategy, to create upward momentum and bring oneself back into a high range. There is no power law distribution in the scores because they are constrained to a small range.

I don't think the "inequalities" really generally apply to constrained reputation economies.



That's kind of like Nigel from Spinal Tap's argument that his amp goes to 11.

I doubt that internally Lyft stores your score on a scale of 1 to 5. It's almost certainly derived from a bunch of other data, or at the very least stored as a floating point number. You can approximate whatever curve you like with either, and then just round.

I'd believe that Lyft doesn't currently have a power law curve, but the 1-5 system isn't what's stopping them.


I know Uber stores as a float. Lyft probably does not since it's averaged over 100 rides. Integer math is easy!

Also you can't mathematically have a power law over (1..5)


You can't have a power law over 1..5? Yes you can. Power law works perfectly fine with finite ranges. It works perfectly fine with discrete values as well. The only question is how accurate a power law model would be in this case.


How about :

100-90% you get a 5

90-75% you get a 4

75-40% you get a 3

40-5% you get a 2

5% and under a 1




Consider applying for YC's Fall 2026 batch! Applications are open till July 27.

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

Search: