Comments on: Shining a Light on the Traffic Signals of Sydney

By: Sam

Sam — Fri, 06 Feb 2026 02:29:58 +0000

The main reason updates to traffic signal programming can take a long time is that it is a manual process. Someone has to go out on site, switch off the signals, install the new program, and switch them back on. This requires traffic management. The traffic management alone can be cost-prohibitive, meaning ad-hoc changes to signals are avoided.

By: Wimal Samarasinghe

Wimal Samarasinghe — Fri, 02 Aug 2024 03:31:09 +0000

Hi Jake,
This is an article about Google’s Project Greenlight to optimise traffic light changes using Google Maps data:
https://sites.research.google/greenlight/

Was wondering if Transport for NSW (TfNSW) might benefit from this information for traffic in Sydney?

regards,
Wimal

By: Jake C

Jake C — Fri, 17 Nov 2023 06:42:10 +0000

In reply to David Nash.

Thanks for this comment. I don’t think I’ve made the claim in this blog post that the average wait time is the flashing red time + steady red time – but if I have I’m very happy to edit and attribute your correction. The `Average max wait` on https://betterintersections.jakecoppinger.com/longest-and-shortest-waits is the average of all samples for a given intersection (calculation function at https://github.com/jakecoppinger/better-intersections/blob/17197acb2a256abd1ac98fbdd2c484cf643077ae/src/utils/utils.ts#L78)

> Max wait times for peds would never exceed one cycle.

Pedestrian wait time can exceed one cycle if they didn’t arrive at the intersection soon enough to press the button, and their cycle was skipped. I believe this is dependant on the model of controller for how long before the potential green pedestrian light pedestrians must press the button to “reserve” their crossing. Again – the insanity at having to apply to cross the road.

> Max wait times for cars and trucks can be up to 3 or 4 cycles in congested conditions.

Yes, I agree, however this post is focusing on delays for pedestrians. A considerable amount of literature covers level of service _for motorists_, however pedestrians are often left out. I don’t believe there is even a concept of LoS for pedestrians.

By: David Nash

David Nash — Tue, 03 Oct 2023 09:55:03 +0000

You seem to be assuming that the average wait time is the flashing red time + steady red time. This is mathematically incorrect. This is the maximum wait time, not the average wait time. Assuming random arrivals, the probability of hitting a red is r/c and the average delay if arriving on red is r/2; so the average wait time is r^2/2c. Traffic Engineering 101! Max wait times for peds would never exceed one cycle. Max wait times for cars and trucks can be up to 3 or 4 cycles in congested conditions.

By: Jake C

Jake C — Wed, 12 Jul 2023 07:11:03 +0000

In reply to espie.

Thanks very much!

Appreciate you finding that – I’ve updated the link.

By: espie

espie — Tue, 11 Jul 2023 08:44:37 +0000

Hey,

I enjoyed reading this, it’s an interesting topic and I like the way you present the data.

Had a poke around the VIC website and looks like they may have updated the link for the public traffic data by adding a hyphen between road and use (road-use). This link looks to work.

https://www.vicroads.vic.gov.au/traffic-and-road-use/road-network-and-performance/road-use-and-performance

This is in regards to the section on your page:
“The Victoria Department of Transport even release their manual count data from detector and survey locations, though the link appears to currently be broken.”

Subscribed!