[Terretta]

> I don’t want to rate my driver. I want to be able to rely on a third party...

You sure can’t rely on the Uber, Lyft, Juno ratings. It’s 5 stars or bust. The social pressure on 5 stars is enormous.

Netflix moved to thumbs up, thumbs down. YouTube did the same, after showing a graph of the 5s and 1s:

https://techcrunch.com/2009/09/22/youtube-comes-to-a-5-star-...

I relentlessly give an average delivery or ride 3 stars, but feel bad every time. When the ride is quite good, 4 stars, and exceptional, 5 stars. Exceptional is the exception.

Three stars doesn’t make you a bad rider or a bad driver, just average. If it’s not the bulk of the ratings you give, you’re an unreliable rater and not helping the ratings anyway.

[dangus]

Your inability to follow the cultural norm is actually at fault here, not the rating system. What you’re doing is akin to tipping 5% in restaurants in the US and only tipping 15% when you have an outstanding experience. In reality, restaurants workers in the USA would expect 15% to 20% unless they dramatically fucked up. I don’t like the tipping system either, but I’m also not going to be that asshole who tries to change the system all on my own without anyone else agreeing to it.

Just use the rating system like everyone else and get over it:

If the driver was great it’s 5 stars with all the “what did I do great” options checked and a note for the driver.

If the driver didn’t fuck up it’s five stars.

If you don’t want to be matched with the same driver again but they didn’t do anything egregious it’s three stars.

If you were outright disgusted at your ride it’s 1 Star.

That’s it. It’s simple. Your own personal usage of the ratings system is not helpful.

Actually, for another example of why your ratings method is bad, let’s compare three stars to grades in school. Three out of five stars would be 60%, which is a D- in most schools. That’s not an average grade. Someone who completes all the homework and does an average job would expect a B, which would be 4 stars. Someone who didn’t get any questions wrong would get an A, 5 stars.

If your Uber driver took you to your destination with a reasonably clean car that’s an A. There’s no such thing as exceptional. It’s a car ride not a physics exam, what do you want exactly?

Uber wants a driver to maintain over a 4 rating, something like 4.5 or 4.2. When you give that driver a 3 rating you’re not saying “thanks, you were acceptable and average.” You are saying “you kind of suck” and Uber won’t actually even match the driver with you again. So if you continue to give all your drivers 3 stars just because you wish the rating system worked a different way than it does, you’re even screwing yourself by reducing the number of drivers that can match with you.

[gjulianm]

From what I have heard from drivers, anything less than 5 stars is bad. Not only with Uber, but with all those companies pushing customers to review their employees. The system is counterintuitive and most people get it wrong at the beginning. Compare what makes you rate 5 stars when you buy a product to your Uber rating system.

However, the think that irks me the most is that rating everyday experiences is just dumb. Most taxi drives will be average and that's it, because we all just want it to be good enough. It's as if my local supermarket made me rate the cashier with 1 to 5 stars. I don't want to do that, because that person just needs to do their job. Anything above "good enough" is unnecessary. Significantly bad experiences should be a "reported to the manager" (or any similar mechanism), filtering out trivial complaints that you'd get in a 5-star scale and getting actual useful information on how to improve the system.

The US restaurant example is funny because the problem is the same. Instead of paying by default fair wages and paying attention to customers that complain about workers, they delegate the 'rating' part to customers, which means that there's no feedback on which they can improve and that their salary is determined by arbitrary people judgements.

[DiogenesKynikos]

> let’s compare three stars to grades in school

That very much depends on which country you're talking about. That's the case in the US, but try telling a teacher in France you deserve 16/20 because you did an average job!

[Izkata]

> What you’re doing is akin to tipping 5% in restaurants in the US and only tipping 15% when you have an outstanding experience. In reality, restaurants workers in the USA would expect 15% to 20% unless they dramatically fucked up.

This is actually the sole reason I don't go to restaurants/diners. These rules aren't what I grew up with (yes, in the US, I've never left the country), seem to be different every time I hear them, and usually keep creeping higher. I just never know what to tip, so it's either fast food and no tip, or go a bit hungry until I can get home.

[dangus]

That’s kind of an unreasonable solution to this problem. I don’t think the 15% rate has changed any in the last multiple decades. I think many people tip 20% because the math is easier.

Tipping doesn’t need to be a source of anxiety. If the act of figuring out your tip is the sole reason you don’t go out for a nice meal on a special occasion, I think this is something to talk to a therapist or confidant about.

[user5994461]

While that's on point, the US school grades are even more obscure and meaningless than the taxi rating to the rest of the world. Many countries are using numbers like out of 10 or 100 and doing an average job on the homework will land you the average like 5 or 50.

It's kinda funny how everything is connected: school grades, restaurant tips, taxi ratings.

[Terretta]

Not inability, choice.

If enough people make the right choice instead of the herd behavior, the commons will be less tragic.

“If the driver didn’t fuck up it’s five stars” is aggressively harmful to any above average or excellent drivers out there, with no reward for trying to be either.

[newswasboring]

Maybe they should scale your ratings based on your rating pattern. I don't know how to do it mathematically but there must be some way to normalize scores if you have enough data. Anybody knows how to do this?

[jankassens]

I’d start with assuming everyone’s experience is normal distributed. Someone who always voted 3 stars probably had the same experience as someone who always voted 5 stars. I’d try out computing the normal distribution of the rater and see where on the distribution this particular vote fell.

[CaptainZapp]

That's exactly what bugs me so much about Trip Advisor ratings. Even then, when they still had some utility.

When you have a 5* scale for rating a restaurant my description would be:

  *      A disaster level lousy place
  **     Sub par. Probably wouldn't visit again
  ***    Quite OK. Probably not my fave anytime soon, but 
         fine
  ****   Above average. Excellent food, service and 
         atmosphere
  *****  An out of this world dining experience. Perfect in 
         every aspect

I realise that there is a certain amount of relativity and subjectivity and that a 5* Trip Advisor review is not necessarily equal to three stars by Guide Michelin.

But it should mean something and when most restaurants have something between 4 and 5 stars (Let alone that the #1 rated restaurant in London was one, which didn't exist[1]) the value of such ratings become very questionable.

You can see the exact same with Airbnb ratings where 4 -, or 5 star does not mean that you will have a great experience.

[1] https://en.wikipedia.org/wiki/The_Shed_at_Dulwich

[ward]

I used to think like this (and still do for movies and such), but recall reading that a driver under 4.6 or 4.7 rating will barely get any clients matched any more.

[ceejayoz]

I desperately wish they’d just adopt an “it was fine” and a “something wrong or exceptional” button.