The algorithm decides WHO gets the good trips!

OldBay

Well-Known Member
Do you really trust the algorithm?

After having a stellar day with multiple long trips, the next day I get stiffed. A few short trips worth $20 total over five hours. Despite getting out early and crawling the spots that give airport runs, nothing!

I am of the opinion that the algorithm uses factors other than location to decide pings.

If you've had good fortune, its guaranteed the algo will balance that out. What if its based on income? You provide great service and get tips, therefore you wont get as many pings! (Need to give Paco in his rusty Sentra some pings.)

Noobs get better/more pings in the first two weeks. People call it the honeymoon. I can vouch for this phenomenon.

What is to say that maybe FT get better pings than PT? Or perhaps long term drivers get better pings than those under a year?

I don't trust the algorithm. Working for an invisible black box that may favor some drivers over others is the pits.
 

DriverMark

Well-Known Member
Which is all the more reason to be on multiple platforms driving. Possible you might get overlooked due to algorithm, but we can only speculate and theorize on that. There are many factors that could explain why no pings. But, if Uber does throttle you for some reason, Lyft will still be pinging.
 

FLKeys

Well-Known Member
One recent evening it was very slow and I get a ping and start heading that way. Pax calls me right away to let me know where they are going because the last 2 drives cancelled on him when they found out he was going 16 miles south, 10 miles out of the hot zone. I took the trip because I knew I could work it for a decent tip thanks to the ants cancelling on him.

Just for the record our rates are higher so long trips are not a loss.

Because I had to pull over and talk to the pax it made the 5 mile pick-up fall into the long pick-up time frame and I got paid for 4 of the 5 miles to pick him up. I have been averaging 2 trips an hour during the week. It is that slow. After dropping off the pax 10 miles out of the hot zone I was rewarded with a stacked ping 7 miles back toward the hot zone, where there were clearly other drivers closer and I still had to drop off my current pax. This also included 4 miles on a long pick up fee. First pax thanked me and tipped $10.

The rest of the night pings came pretty quick with less than 5 minutes between them. All the time passing other drivers.

Did the algorithm reward me for taking the trip others cancelled on? One will never know.

I had an outstanding Monday & Tuesday this week, will the algorithm punish me tonight? How will I really know, maybe it was just an off night.
 

Rakos

Well-Known Member
Try praying to the Uber god's...

Sometimes the Uber god's reward me...

Sometimes the Uber god's don't...8>)

Butt...when they do hang on...!!!

Rakos
monkey-scratching-head.jpg
 

Soldiering

Well-Known Member
I've been a driver in PHX for almost 3 years. I'm sure the algo does have certain switches for lack of a better term and In my experience you can lose your place so to speak. I drive both platforms an thats why if one craps on me I've got the other. I accept an cancel rides whenever a closer pops up or the drive to is congested. I've never been overtly punished for this and I do 100 to 150 trips a week. My two ¢............Stay cool out there fellas
 

touberornottouber

Well-Known Member
I've been saying this for years. I actually have experience designing algorithms in a somewhat related application years back. I can see it in action. Sometimes it is very obvious. I will often have days where I get 20 Lyft trips and only 2 Uber trips. I'll typically see some very bizarre repetitive patterns such as where I will go months having 3/4 weeks during a month falling between $100 - $130. Meanwhile Lyft will earn me between $300-$500 while on at the same times. I've also noticed I will usually have one good day with Uber and then all the other days will be between $10-$30 -- even if I stay on for 8-10 hours. etc. etc. I could go on a lot. There is definite funny business.

And don't forget there are even patents from Uber available in full view of the public where they detail how they discriminate against drivers in other ways besides just the ETA. See http://www.freepatentsonline.com/y2017/0011324.html
 
Last edited:

Boca Ratman

Well-Known Member
lol, drtivers can't even agree on what is a "good trip" and you think some computer program not only knows what you consider a "good trip" but allocates them in some order?
 

NotanEmployee

Well-Known Member
It's a lot to read and seems like repetitive nonsense but I find it interesting that it matches the rating the pax gave you to both the pickup and destination. This suggests the alogrythm knows certain neighborhoods result in lower ratings for both pax and drivers. The biggest problem here is I'm not sure it's using that info to both the drivers and the pax benefit. It really could block out bad locations from service.......if it wanted.....but uber will risk it b E cause they don't have to drive there.
Post automatically merged:

I would love to set a home area...this area would be a 5-10 mile radius or so. Any time you travel outside the radius it tries to get you back in it. I'm safe living on an island but can see the nightmare of a Reno driver taking a pax to Sacramento, then getting one to San Francisco, then to San Jose...etc. Worst nightmare. Instead Sacramento is as far as he goes. All future rides head in the direction of Reno until he has to dead head back over the summit.
 

OldBay

Well-Known Member
  • Thread Starter Thread Starter
  • #13
There is a small parking lot in the city where me and two other drivers sit. I noticed that it implements a queue (first in, first out) for pings.

These guys were sitting there together before I started going, which tells me it doesn't matter where we are in the are, if drivers are in the vicinity, it does some sort of "fairness" algo. If there wasn't this "fairness" build in, they both would have camped in separate areas of the city. Also, if they were parked marginally closer they would get the pings, but it doesn't seem to work that way.

I suspect when it gives you pings, you go colder. Those who haven't had pings are hotter. Everyone in the area is graded and those who are the hottest get the pings first. Doing things like canceling rides makes you colder.

This "temperature rating" may span multiple days/weeks/months. Unclear how much relative distance has to do with getting pings. I suspect within a certain distance (say 4 miles) it doesn't matter as much as how hot a driver is.

When you are on the clock and not getting pings, that raises your temperature.

I think what is happening is that I have been working hard, getting pings, and then some part timers join the fray, haven't been driving and are hotter, and they suck up the pings.
 

TomTheAnt

Well-Known Member
I definitely pissed off The Algorithm when I put it on a hiatus for four weeks (only took one ride) and only drove Lyft. It just hasn’t been the same ever since. And Lyft keeps throwing good bones, and I keep taking them.
 

BigRedDriver

Well-Known Member
I've been saying this for years. I actually have experience designing algorithms in a somewhat related application years back. I can see it in action. Sometimes it is very obvious. I will often have days where I get 20 Lyft trips and only 2 Uber trips. I'll typically see some very bizarre repetitive patterns such as where I will go months having 3/4 weeks during a month falling between $100 - $130. Meanwhile Lyft will earn me between $300-$500 while on at the same times. I've also noticed I will usually have one good day with Uber and then all the other days will be between $10-$30 -- even if I stay on for 8-10 hours. etc. etc. I could go on a lot. There is definite funny business.

And don't forget there are even patents from Uber available in full view of the public where they detail how they discriminate against drivers in other ways besides just the ETA. See http://www.freepatentsonline.com/y2017/0011324.html
I enjoyed your post very much. I’ve noticed weird patterns. One day I’ll pick up or drop off 5 or 6 times in a single day, at an apartment complex, then not again for a month.

Or I’ll go to pickup down a specific road 3 or 4 times in a single day, then not hit that road again in a month.

It’s almost like the algorithm is trying to get me use to an area, then move me to the next so I get good at all?

Makes my head hurt wondering why this chit happen?

One week I’ll get called to a high school 6 or 7 times, then not again for weeks?

Huh?
 

goneubering

Well-Known Member
Do you really trust the algorithm?

After having a stellar day with multiple long trips, the next day I get stiffed. A few short trips worth $20 total over five hours. Despite getting out early and crawling the spots that give airport runs, nothing!

I am of the opinion that the algorithm uses factors other than location to decide pings.

If you've had good fortune, its guaranteed the algo will balance that out. What if its based on income? You provide great service and get tips, therefore you wont get as many pings! (Need to give Paco in his rusty Sentra some pings.)

Noobs get better/more pings in the first two weeks. People call it the honeymoon. I can vouch for this phenomenon.

What is to say that maybe FT get better pings than PT? Or perhaps long term drivers get better pings than those under a year?

I don't trust the algorithm. Working for an invisible black box that may favor some drivers over others is the pits.
You’re right. Uber’s algorithm uses many factors.
 
Top