Sorry I'm stupid – I meant L1 regression! Not sure why this tweet blew up but I might delete and repost
I have had a weird theory for years that Uber fits their ETA model using Lasso instead of ordinary least squares. Which causes a bias towards shorter wait times.
I have had a weird theory for years that Uber fits their ETA model using mean absolute error regression instead of ordinary least squares. Since wait-times have a long right tail, this will cause a bias that underestimates mean wait times (median < mean).
Standard experience booking an Uber
“5 minutes away”
I order it
“Finding your driver…”
“Pickup in 7 minutes”
I open my timer
9 minutes and 21 seconds later, it arrives
This is so tedious and I am tired of being deceived at the margin
I have had a weird theory for years that Uber fits their ETA model using L1 regression instead of ordinary least squares. Since wait-times have a long right tail, L1 reg will cause a bias that underestimates mean wait times.
Standard experience booking an Uber
“5 minutes away”
I order it
“Finding your driver…”
“Pickup in 7 minutes”
I open my timer
9 minutes and 21 seconds later, it arrives
This is so tedious and I am tired of being deceived at the margin
@bernhardsson loss functions quietly become product behavior in cases like this
@bernhardsson no!! where is the original post!?
@bernhardsson For me it says 2 minutes away when the driver is literally right in front of me
Sorry I'm stupid – I meant L1 regression! Not sure why this tweet blew up but I might delete and repost
I have had a weird theory for years that Uber fits their ETA model using Lasso instead of ordinary least squares. Which causes a bias towards shorter wait times.