Has Anyone Ever Tried to Algorithm Uber and Lyft Back?

Alloverthemap

Active Member
I think it's conventional wisdom by now, that Lyft -- and maybe Uber -- tries to "even out" drivers' earnings in order to eliminate fluke great days and fluke horrible days. The fluke horrible days could dissuade drivers from continuing to drive with them, while the fluke great days aren't likely to attract more drivers.

So, if you make $30 in one hour, I think your odds are artificially low of making $30 in the next hour, whereas had you not had that initial $30 strike, the odds would be greater.

But we still have the power to accept rides when we please. If I leave the app on long enough without accepting a ride I will see the same names start popping up over and over again. To me this suggests that we're not being shielded from any potential rides. But maybe their algorithm will shield us from those good rides popping up first after we've had a good run of rides for an hour.

Let's say there are eight people awaiting rides with no drivers biting. It would seem likely, if U/L are averaging things out per driver, that these fares would be ranked in order of how they want to assign them to each driver. First ping: $4 ride, second $4.15, eighth $19 . . . see where I'm going with this?

Are we aware of anyone actually tracking this? That is: Average fare of first ping through eighth ping (or beyond)? If they're playing averages against us, it seems to me that we could figure out how they do it and play them right back against Uber and Lyft.

Thoughts?
 

SHalester

Well-Known Member
....wouldn't this only work if for every ping you get there were several other drivers around who may or may not have received the same ping and declined it for various reasons?

too many variables involved. Pretty sure the system doesn't go eenie, meany, minny moe who made too much (or not enough) $$ here you go have a ping.
 

Alloverthemap

Active Member
If what you say is true, then we need not worry that just because we got a great ride early in the day, the rest of the rides will bring us down to some sort of average. The remainder of those rides should approximate our daily average, but not be biased to go below it.

And I hope you're right!
 

SHalester

Well-Known Member
And I hope you're right!
there are a lot of variables involved. How many ants around you is a big factor; maybe the biggest. Somehow the system has to break 'tie-breakers' and award the ping.
One variable that occasionally comes up here: eye color. That is a day when a bunch of posting members are very bored. Not bleepin way is eye color considered, right? 🤷‍♂️ :eek::whistling:
 

rushbudgie

Well-Known Member
Basically the wealth is distributed to say $30ph. The newcomers are given much more rides enticing them to stay. Soon their honeymoon is complete and join the rest of us. Prime example: I had a pax and received a stack ping. We were stopped at traffic lights when this occurred, the strange thing is opposite me another uber driver (could see his phone) did not receive the ping. Probably because he was at $30ph and I wasn't so Uber gave it to me to reach the golden $ph.
 

Boca Ratman

Well-Known Member
just because we got a great ride early in the day,

What's a great ride?

Ask 10 people here and you’re likely to get 10 different answers.

The last tbink uber cares about is spreading the earnings out. Drive enough and you'll have great days, shitty days and a bunch of average days. I've had days where I made $500 and had maybe 15 dead miles . And I've had days where I had to dead head home 150 miles.

Its all about getting a ride to the pax the quickest. You as a driver only factor in to that equation as a possible solution.
 

Alloverthemap

Active Member
Basically the wealth is distributed to say $30ph. The newcomers are given much more rides enticing them to stay. Soon their honeymoon is complete and join the rest of us. Prime example: I had a pax and received a stack ping. We were stopped at traffic lights when this occurred, the strange thing is opposite me another uber driver (could see his phone) did not receive the ping. Probably because he was at $30ph and I wasn't so Uber gave it to me to reach the golden $ph.


Agree about newbies. When I switched over to Lyft, it was a succession of $10 rides for that first couple of days. They did coincide with colleges opening, so I wasn't convinced about the sugarplums. Sure enough things fell off radically from there, though still competitively with Uber since I was always in motion with or towards a pax. With Uber there were long dry spells on occasion, though the rides were typically longer.

So, the second part of your answer supports my suspicion that there is a regression to the mean algorithm at work in assigning rides. Therefore, I think a study would be valid to see whether it's smarter to wait for that, say, fifth ping before pouncing than what your normal system would advise.
Post automatically merged:

What's a great ride?

Ask 10 people here and you’re likely to get 10 different answers.

The last tbink uber cares about is spreading the earnings out. Drive enough and you'll have great days, shitty days and a bunch of average days. I've had days where I made $500 and had maybe 15 dead miles . And I've had days where I had to dead head home 150 miles.

Its all about getting a ride to the pax the quickest. You as a driver only factor in to that equation as a possible solution.

If this is so, then all rides are assigned randomly. Which I hope is the case.
 

rushbudgie

Well-Known Member
Agree about newbies. When I switched over to Lyft, it was a succession of $10 rides for that first couple of days. They did coincide with colleges opening, so I wasn't convinced about the sugarplums. Sure enough things fell off radically from there, though still competitively with Uber since I was always in motion with or towards a pax. With Uber there were long dry spells on occasion, though the rides were typically longer.

So, the second part of your answer supports my suspicion that there is a regression to the mean algorithm at work in assigning rides. Therefore, I think a study would be valid to see whether it's smarter to wait for that, say, fifth ping before pouncing than what your normal system would advise.
Post automatically merged:



If this is so, then all rides are assigned randomly. Which I hope is the case.
I take every ride so that my hourly rate is constant so if I was to wait for that fifth ping Uber may lessen my pings. I'm letting Uber know I want those pings short or long, nothing worse than waiting & driving dead miles. It's luck as well, that's my guess.
 

I will crack Lyft hacks

Well-Known Member
Thoughts?
Interesting thoughts.

Of course our income is played with through algo.

The dispatch algo is black boxed, meaning we will never know how it works. Otherwise we would manipulate it, not it manipulating us.

Ride matching takes into factor how long you have been without a ride. If you have been waiting 30 minutes and your Per engaged hour income is falling below average, the algo will choose you over the closest driver.

I always try to check when the order comes in to see if I’m the closest driver. A lot of times that is not the case. Proving that proximity to the pax is only 1 factor.
 

Fusion_LUser

Well-Known Member
I always try to check when the order comes in to see if I’m the closest driver. A lot of times that is not the case. Proving that proximity to the pax is only 1 factor.

Yeah I'm seeing that proximity factors very little and with the rides I take at 2.5x or higher I will get a pickup request 15 minutes away and when checking the pax app I see 1-3 ants closer. Usually those rides are minimum fare rides which suggest the others ants declined them. But then there are some that are usually good rides for any ant that still come to me with my surge rates while other ants are closer.

The only thing I'm seeing consistently is pax thanking me for picking them, that they had a hard time getting a ride. I could see 5-6 ants on the pax app but the pax are saying nobody would take them but me. Doesn't matter how far the pickup or the surge rates either... the pax seem happy to just be able to get a ride.
 

Uberguyken

Well-Known Member
Gotta work the surge game...

Screenshot_20200919-115500.jpg
 
Top