Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I once read (sorry can't find cite) that Waze and Google maps run roughly the same routing algorithm now, but Waze is hyper-optimized for time whereas Google maps is optimized for some nebulous quality akin to "fastest reasonable route" where unreasonable factors are things like using small roads, exiting and entering the high a bunch, extra turns, etc.

So Waze probably is faster by a bit, but not a whole ton.

In my experience, Google Maps ETA has gotten to be frighteningly accurate in the past 6 months. For situations where traffic doesn't materially change during the trip, my ETA will be within 2 minutes of arrival even on trips of over an hour.



I've noticed Waze optimizing for quickness. Not fastest, not shortest, not most direct but the route which makes you feel like you are getting there quicker. Often longer stretches with no traffic or junctions for example.

It's as if users main priority is not to be held up. Users don't care if the journey is two minutes longer or a few pence more expensive, they just don't want to be feeling they are making no progress.


" they just don't want to be feeling they are making no progress."

I think you may have just convinced me to use Waze instead of Google Maps. I'd much rather relieve the stress of driving a bit than save a few minutes of time.

Well, actually, I'd rather not drive at all. But seeing as that isn't really an option, I'll settle for making it less stressful.


I’m 36 and have lived in a mix of rural and urban areas of America and I never learned to drive and have never driven a car.

Just want to make the point that not driving is really an option, just one that most people look for a way to avoid. It’s all a matter of priorities and as much as most people say they hate driving/traffic, very few actually do anything about it.


IME Waze is actually dangerous because it doesn't mind sending you on crazy maneuvers like making uncontrolled left turns across 8 lane roads only to immediately merge across three lanes and hang another turn. At times it saves a couple minutes, other times it backfires spectacularly, and at all times it disdains safety and simplicity. For me, it's much higher stress.


At some level that's true. But it's also true it significantly constrains your life choices. I've had one job in my life that I could maybe have reasonably held without being able to drive. (Business travel would still have been somewhat of an issue.) And my entire personal lifestyle would have to have been totally different.

Also, a lot of people just aren't in a position to realistically say "no" to any job that involves a need to drive.


I think it's a little ridiculous to say "I would rather not have to drive" and then make 0 life choices to reduce the reliance on driving, which is exactly what most people do.


People do often take the length of commutes into account. But, while many would prefer not to have to drive in the abstract, in practice that preference doesn't rise to the level of:

- Moving into a smaller place in a city

- Potentially putting kids into worse schools

- Passing on many of the recreational and travel activities that require driving

- Passing on job opportunities that have better salaries, career paths, etc.

- Not visiting friends and family who live places you need to drive to

All of which may be the result of not being able to drive or making a decision to prioritize minimal driving over everything else. There are a lot of things that people prefer in principle but, when presented with the actual choices, they make tradeoffs that favor different things.


It's true that I could make choices that would reduce my driving further. I could Uber, or take local buses. I could even ask my wife to drive me around, and she'd probably do it, bless her heart.

But Uber is way more expensive, local buses would add hours to my commute every single day, and I'm not nearly selfish enough to make my wife drive me to and from work every day.

Oh, and bicycling is far, far too dangerous here. Otherwise, I'd totally pick that option.


Which rural areas have you lived in?


> It's as if users main priority is not to be held up

I believe this to be a very real and widespread emotional phenomenon. Not only have I observed it in countless other drivers, even as a quite experienced driver myself, I still feel it on a regular basis. Sitting at a stoplight for too long sometimes just makes me irrationally angry, and I will often add a couple minutes to my commute to avoid the experience. I know it's stupid and am trying to change the behavior, but it's difficult.


The psychological pressures of 'progress' and 'entrapment' (E.G. being trapped in gridlock or long lines) are not to be underestimated.

If rush hour were still actually an _hour_ rather than a smear that covers all of daylight hours and then some, happy hours and other 'avoid the traffic' distractions would likely see more utilization as an alternate means of avoiding the above problems.


I have had Waze give me some very good detours, for example driving to northern Michigan on a holiday weekend where there's a ton of traffic, get off the freeway before the massive traffic jam and onto the parallel 55mph highway, then get back on once the jam is cleared.

OTOH, it was responsible for getting me my only traffic ticket, for violating a (very hard to spot) no through traffic sign.


Do you use turn by turn navigation in Maps? It could've adjusted its predictions to your driving patterns.


I'd assume any kind of "ETAnet" would be trained on everyone's driving patterns, not individuals.


I don't know, but I always assumed that a simple optimization would be one that compared your typical driving speed to posted speed limits on open roads. That is, do you usually drive 5 or 10 miles (kilometers) per hour under or over the limit when you have the choice?


I suppose they could. I just think there's enough driving patterns that can be detected in aggregate, that a network could detect your typical style much better with everyone's data, rather than training just from you.


There must be some training involved on the user level. My route to the freeway can take 1 of 2 parallel roads. The bigger road is theoretically a minute faster but often enough the left turn lane backs up for multiple light cycles. The smaller road is quite consistent in it's timing. A minute slower if the lights work out and a minute faster if you would have to wait.

I prefer the consistent route and now Google almost never tries to send me the "faster" way.


My experience with Google Maps is that the accuracy is proportional to how close you are to Mountain View.


It's very accurate in Budapest so I don't think thats true.


In my experience the walking directions are the ones that suffer the most. Here's an example I discovered on my last vacation that still looks like it happens. Try asking Google Maps for walking directions to the Arc de Triomphe in Paris from any point nearby such as the George V metro stop. For anyone unfamiliar with how to get there, there is a pedestrian tunnel from the Champs-Elysees that goes underneath the roundabout and comes out right at the Arc de Triomphe.


In the Twin Cities as well.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: