Signal timing rations the day and pedestrians get the remainder

There is a settled way of talking about traffic signal timing. It is worth asking how much of it survives contact with the detail.

The cycle is a fixed budget

A signalised junction runs a repeating cycle, and every movement gets a share of it including the clearance time between phases. Lengthening the cycle raises vehicle throughput because proportionally less time is lost to changeovers.

It also raises the maximum wait for anyone not currently being served, and pedestrians are usually the movement served least often. A junction optimised for vehicle capacity therefore produces long pedestrian waits as a direct arithmetic consequence, not as an oversight.

Pedestrian stages compete with turns

Where a pedestrian crossing conflicts with a turning movement, the two cannot run together without either a filter arrangement or a shared risk. Giving pedestrians a protected stage removes that turn from the cycle for its duration, which is a measurable capacity cost.

This is the single most common reason a crossing is omitted on one arm of a junction, leaving the notorious three-sided crossing. The pedestrian is then required to make two or three crossings to reach a corner that is a few metres away.

Walking speed assumptions decide crossing time

Green man duration and clearance time are calculated from an assumed walking speed applied over the crossing length. Assumptions drawn from average adults leave a significant share of older and disabled pedestrians short of time. Several cities have revised their assumed speeds downward after review, and doing so lengthens the cycle or reduces vehicle green.

Measured properly, that trade is the reason the revision is contentious rather than routine.

Waiting produces informal crossing

Pedestrian compliance falls sharply once the wait exceeds roughly half a minute, and the relationship is well established in behavioural studies even if the exact threshold varies. The response in many places is enforcement or barriers rather than shorter waits, which treats a design signal as a discipline problem. Guardrail installed to prevent crossing also traps pedestrians in the carriageway and obstructs the footway, so it carries its own costs.

Shorter cycles with more frequent short pedestrian stages generally produce better compliance than long cycles with generous stages.

Detection is asymmetric

Vehicle detection through loops, radar or cameras is standard and continuous, so the system knows exactly how many cars are waiting. Pedestrian demand is usually registered by a push button, which records that at least one person is present and nothing else.

Adaptive systems therefore optimise against a rich vehicle dataset and a binary pedestrian one, which biases the outcome regardless of intent. Cities deploying pedestrian detection have been able to extend crossing time when a slow user is still in the road.

What works in a dense grid may not transfer to a low-density suburb.

Removing signals is sometimes the answer

At junctions with modest vehicle flows, signals impose delay on everyone in order to manage a conflict that geometry could resolve. Replacing them with a compact roundabout, a raised table or a simple priority arrangement can reduce total delay.

Across a network, the evidence on shared-space schemes that remove control entirely is genuinely mixed, and blind and partially sighted people have raised well-founded objections to the loss of controlled crossings. Signal removal is a tool with a narrow band of appropriate use, not a general principle.

The takeaway

A crossing wait is not an accident of technology. It is a number somebody chose.

Cities are built by a thousand small permissions, not one big plan.