Changes that would be appreciated!

[CBenson]

+1

[condrek]

It's the average over the entire length of the segment. It IS very accurate except when conditions are changing rapidly due to something that has just happened. If it weren't accurate my ETAs at the start of my ~25 mile daily commute wouldn't be +/- 1 minute of when I actually arrive.

Your ETA's may be accurate where you are living, but the average speed that waze reports in my area (NYC) are not. My ETA's are constantly fluctuating... That would be the reason for my suggestion.

I have zero faith that the average Wazer can accurately estimate the speed of a longer traffic jam. Human vs computer (with well written software). Which is generally more accurate with mathmatical calculations?

Oh, ye of little "faith"... The computers can still do the calculations, just using the feedback that the user provides as well as the road data. Besides, how hard is it really to average out your speed in a traffic jam? I would say it's pretty easy when you're crawling along a roadway.

[bgodette]

Oh, ye of little "faith"... The computers can still do the calculations, just using the feedback that the user provides as well as the road data. Besides, how hard is it really to average out your speed in a traffic jam? I would say it's pretty easy when you're crawling along a roadway.

Which it already does and it already displays. What you suggested in the OP in regards to jams already exists; a jam will end on the last segment that has abnormally, compared against historical, slow transit. It's pointless to display it ending mid-segment since you can't join back in mid-segment unless your car happens to be equipped with an oscillation overthruster to phase through those buildings in your way.

[CBenson]

I don't believe that the jam display always displays across whole segments if the traffic data doesn't show the entire segment to be slow.

[jasonh300]

I don't believe that the jam display always displays across whole segments if the traffic data doesn't show the entire segment to be slow.

Traffic data is average speed calculated across the segment. It would have to show the whole segment.


I'm not drunk...I'm using Tapatalk.

[CBenson]

I don't think it is average speed across the segment.

The question of whether long freeway segments should be broken into smaller segments so that waze will more accurately pinpoint jams has been frequently raised. My understanding is that waze's position is that it is not necessary or helpful.

I am also a reasonably certain that you do not need to traverse an entire segment for waze to detect the jam. I have been on freeway segments that have been completely blocked by an accident. No one who was delayed got by, yet waze would indicate the jam within minutes of the accident.

[gettingthere]

I am also a reasonably certain that you do not need to traverse an entire segment for waze to detect the jam. I have been on freeway segments that have been completely blocked by an accident. No one who was delayed got by, yet waze would indicate the jam within minutes of the accident.

Yes, this is correct. The entire segment is updated with the jam speed and the speed will continue to update on this segment until you complete the segment or other Wazer(s) drive the segment.

[bgodette]

I am also a reasonably certain that you do not need to traverse an entire segment for waze to detect the jam. I have been on freeway segments that have been completely blocked by an accident. No one who was delayed got by, yet waze would indicate the jam within minutes of the accident.

Yes, this is correct. The entire segment is updated with the jam speed and the speed will continue to update on this segment until you complete the segment or other Wazer(s) drive the segment.

To clarify how it's able to do this: Waze knows your current position, how fast you're currently moving, how much farther you need to go to exit the segment, and what your average speed has been for the previous portion of the current segment. Given that information it can extrapolate an average speed for the entire segment without you having yet exited the segment. Once a wazer has traversed the entire segment, their now true non-extrapolated average can update the segments real-time average speed.

This means that it is possible for a situation where an accident has happened early in a very long segment and you're the only wazer currently on it; given a long enough delay you'll bring the expected average way down until you clear the accident, then it'll start to climb back up, creating a "flashing" traffic jam. I'd expect this to only happen in low density areas.