It's interesting that there is a cut-off around 4.70 for some drivers, it seems bizarre then that I have a 4.69 rating. I considered myself the ideal passenger. I'm talkative or quiet to match the driver's personality, always polite, shower before the ride, well dressed, never complained to or made any kind of request from the driver, nearly always rate the driver 5/5. Maybe it's because I don't tip? It's very strange to imagine that I'd be getting automatically declined based on my rating.

The cutoff probably varies per city, but I imagine they're only trying to get rid of the lowest percentiles of the distribution.

Honestly though, a 4.7/5 generally sounds good. If a 4.7/5 is near the (very) bottom of the distribution then it sounds like the rating system is quite flawed and should suggest to Uber that they need to replace it (perhaps with just a "would you use this driver again" question, and if no, then whether the cause was behavioural (e.g. odour or some form of rudeness) or dangerous driving), then rank drivers using this. Vice versa for riders.

If the /5 rating system was accurate, then 'good' drivers/riders should be ~3.5-4, excellent drivers 4.2+, and poor drivers 1-2. But since some people have different definitions of 5/5 compared to others the rating system ends up generally turning into a 5/5 for above satisfactory with no real issues, and random usage of the rest of the scale.

Well, this is in my city, I don't know if drivers elsewhere have this kind of informal rule too. This means, more or less, that for 1 ride with 4 stars you need 3 rides with 5 stars.

Or, even worst, that you'll need 13 5-stars to compensate for 1 ride with 1 star...

It’s because you don’t tip. Why don’t you tip anyway?

That happens because of rating inflation. If everyone just used the system where 3 was the average rating instead of 5, then he'd be in a much lower range.

But since everyone got in their heads that anything lower than a 5 is somehow an assault on the other party, 5s are the "average" rating and there is no way to distinguish between average and good. You only know that he was bad enough to get dinged .3 stars over the course of his history with the app.

The issue with this rating inflation is that it's a recursive effect.

If the majority end up treating 5 as average, then you also have to do the same and go with how the majority ends up using the rating system, which in turn forces even more people to use that definition of ratings, and so the cycle continues.

If you hold 3/5 as your average when 5/5 is held as the average for most people, you're just hitting the driver with a "this was awful" effectively.

This rating scale allows for some form of categorisation of the levels of 'badness' of the drive (e.g. differentiating between dangerous driving and jerky driving, between rudeness and general discomfort), but doesn't allow 'above satisfactory' to be separated from good or excellent.

But perhaps that's what Uber is going for with the rating system; maybe their intent is that 5 is meant to be 'above satisfactory' and anything better than that should be a 5 + a tip? If that's not their intent then I have no clue why they haven't redone the rating system already.

Who tells everyone to use that scale though? You're presented with a row of stars. Expecting people to read winds up being a big ask.

Also, those categories are subjective as well. What's the difference between "bad" and "really bad"?

Mutual graded rating systems suck. That's all there is to it. Because no matter what, it turns into currency. "Yes/No" and then base a rating on that. Rotten Tomatoes had it right from the start.

I'd probably flat out rate these annoying passengers with one star or maybe even tell them to get a hike.