> unless you miss the light

I've been running into this constantly. It seems like it has no allowance for the average number of seconds spent at a light, so it sends me routes cutting through town that end up being slower because of all the signals, when it could have just sent me up the freeway three more exits.

I've learned to be critical of the route and look before I leave.

I live in an area that has many traffic circles and Google Maps seems to actively avoid lights and prefer routes that use circles (which is completely fine with me and preferred after I got comfortable using circles). I think Google knows how much time people topically spend going through each intersection. But it may just be the eco-friendly option.

One think that really messed things up is that route option state is pretty hidden. My wife was ranting at how stupid Google Maps was because my daughter was using Apple Maps in the drivers seat and Apple Maps was doing sane things and Google Maps was doing dumb things. Apparently this had gone on for weeks. Turns out somehow she had "avoid highways" selected and didn't realize it.

The app really does seem to get worse and worse though. I think it's trying to push me to switch to using voice commands but I'm too dumb to remember the incantations yet.

I have felt for a while now that Google Maps does a lot of A/B testing on routes. Probably they inherited this from Waze -- I quit using Waze years ago because it routinely picked a route that I knew would be slower from my experience in the area. So I'd ignore the turn it wanted to make in that direction, and it would re-route on the better path... and the time estimate would drop a few minutes. So it -knew- the correct path was shorter but was sending me the longer way just because it wanted to be sure it was in fact the longer way?

Sometimes I wonder if Google engineers wrote this stuff to get users to test different routes.

Randomly screwing over people to make sure the algorithm isn't stuck in a local maximum seems on brand for Google, but I think it's because most of the data comes from Amazon delivery vans, taxis, Uber/Lyft and other professional (and hence, aggressive) drivers.