Traffic calming works when it changes geometry, not when it adds signs

This is less a set of instructions about traffic calming than an argument, and it is worth saying so at the start.

Deflection is the working principle

A driver slows for a hump because passing it at speed is physically unpleasant and potentially damaging to the vehicle. A driver slows for a chicane because the path requires steering input that is uncomfortable above a certain speed.

Both work on the vehicle, not on the driver's intentions, which is why compliance does not decay over time. Anything that relies on the driver choosing to comply — signs, painted roundels, digital speed displays — decays as familiarity grows.

Vertical deflection is effective and unloved

Road humps, cushions and raised tables produce large and durable speed reductions where they are spaced closely enough. Spacing is the usual failure: drivers accelerate hard between widely spaced humps and the average speed barely changes.

At street level, cushions were introduced to let buses and emergency vehicles straddle them, and they are correspondingly less effective on cars that can also straddle them. Raised tables at junctions and crossings combine speed reduction with a pedestrian benefit and attract fewer objections than mid-block humps.

Horizontal deflection buys space

Build-outs, chicanes and pinch points narrow the effective carriageway and force a lateral movement. Because they take space rather than adding an object, they can host planting, seating, cycle parking or drainage.

Across a network, poorly designed pinch points create a conflict where a cyclist and an overtaking driver arrive at the narrow section together. Designs that either give cyclists a bypass or make the narrowing wide enough not to be shared avoid this entirely.

Area-wide beats street-by-street

Calming one street reliably pushes traffic onto the next one, which produces the familiar sequence of neighbours petitioning in turn. Area-wide schemes that treat a whole cell bounded by main roads avoid this and are more expensive to deliver at once. They also allow the internal streets to be filtered rather than simply slowed, which addresses volume as well as speed.

Evaluations of area-wide low-traffic schemes have generally found internal traffic reductions with mixed and smaller effects on boundary roads.

The emergency access argument deserves a real answer

Fire and ambulance services report that vertical deflection slows appliances and can be uncomfortable for patients being transported. That is true and measurable, and it has to be set against a reduction in the collisions those services attend.

Over a decade, net effects are genuinely difficult to establish, and studies have come to different conclusions depending on what they counted. Practical compromises route emergency vehicles along a treated hierarchy and use tables and cushions rather than full-width humps on those routes.

Twenty is a limit, not a scheme

Area-wide lower speed limits have been introduced in many cities, sometimes with physical measures and sometimes without. Where the street already looked like a slow street, compliance is high; where it does not, the limit alone shifts speeds only slightly. Even small average reductions matter because pedestrian injury severity rises steeply with impact speed, so the intervention is worth making.

Over a decade, the honest description is that the limit sets the target and the geometry decides whether it is met.

The takeaway

If a driver could ignore the measure without noticing, it is not a traffic calming measure.

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