Near-miss feedback cuts decision time by 0.6 seconds

A signalised intersection on Warsaw's Marszałkowska Street handles roughly 2,400 vehicles per hour at peak. Each driver approaching the stop line makes a go/no-go call in under two seconds, often with incomplete information about the car ahead, the pedestrian countdown, and the light phase. Transport engineers have long modelled that decision as a function of gap acceptance and reaction time. But a growing body of behavioural research suggests something else is happening too: the feedback a driver receives about how close they came to clearing the intersection shapes how fast they decide next time. If near-miss information measurably shortens decision latency — one recent experimental result puts the effect at around 0.6 seconds — what does that mean for how we design junctions, signals, and driver-assistance systems in Poland?

The near-miss as a feedback signal, not just an error

In road safety analysis, a near miss is usually treated as a failure that didn't quite happen — a data point for conflict studies, counted by traffic cameras and scored by severity indices. Behavioural psychology sees it differently. A near miss is informative feedback: it tells the decision-maker that their estimate of the situation was close but not correct, and it does so with an emotional charge that a clean success or a clean failure doesn't carry.

This is where the classic reinforcement literature becomes relevant. Variable-ratio schedules — where a reward arrives after an unpredictable number of attempts — produce persistent, fast-responding behaviour. B.F. Skinner's work on intermittent reinforcement established that unpredictability itself sustains engagement. Near misses sit in a related space: they are ambiguous outcomes that feel like progress, and they tend to accelerate the next attempt rather than slow it down.

The 0.6-second figure in the headline is the kind of number that shows up when you operationalise this in a lab. Studies of feedback timing in sequential choice tasks — including work on "near-miss" framing in decision experiments — have found that trial-to-trial decision latency drops when participants receive graded feedback about proximity to a target rather than binary win/lose feedback. A 0.6-second reduction is large in absolute terms; in a driving context, where typical perception-reaction time is often cited at 1.0–1.5 seconds for unexpected events, shaving six-tenths of a second off a repeat decision is the difference between stopping comfortably and stopping hard.

What faster decisions actually buy — and cost

Speed is not automatically good. The engineering literature on intersection safety treats quick decisions as valuable only when they're also accurate. A driver who decides 0.6 seconds faster but 10% more often wrongly is a worse driver.

Here the psychology of risk-taking under uncertainty matters. Daniel Kahneman and Amos Tversky's work on loss aversion showed that people weigh losses more heavily than equivalent gains — a asymmetry that should, in theory, make drivers conservative at yellow lights. But near-miss feedback can recalibrate that asymmetry. If a driver repeatedly just clears an amber light, the perceived cost of stopping (time lost, rear-end risk) starts to feel larger relative to the perceived cost of proceeding. The decision gets faster because the internal threshold has shifted, not because the driver has become better at estimating.

This is the double edge. In competitive contexts — sport, trading, even competitive video gaming — near-miss feedback is known to increase persistence and risk appetite. The same mechanism that sharpens a tennis player's return timing can push a driver to treat a shrinking gap as an invitation. Polish road-safety data already flags red-light running and aggressive amber-light behaviour as persistent problems at high-volume urban junctions; anything that systematically shortens decision latency without improving accuracy deserves scrutiny.

Designing infrastructure that gives honest feedback

The practical question for Polish transport planners is whether junction design can supply feedback that is calibrated — informative about real risk rather than merely about proximity to a threshold.

Consider the countdown timers now common on pedestrian crossings in many Polish cities. These are, functionally, a continuous near-miss display: they tell you how close you are to the moment the phase ends. Early evaluations of pedestrian countdown signals in various countries found modest reductions in crashes and clearer crossing behaviour, partly because the timer converts an ambiguous "will I make it?" into a graded estimate. The feedback is honest: it reflects the actual remaining time, not a manufactured sense of proximity.

Vehicle-side systems can do the same or the opposite. A forward-collision warning that fires only at the last moment is a near-miss generator; one that gives graded, early, and consistent feedback about closing distance is a calibration tool. The difference matters for how drivers learn. Research on feedback in safety-critical tasks consistently finds that immediate, specific, and graded feedback improves skill acquisition, while delayed or binary feedback produces either overconfidence or neglect.

There is a concrete precedent worth citing: the Swedish "Vision Zero" approach to intersection redesign leaned heavily on separating conflict points rather than relying on driver judgement. The philosophy there is that human decision-making under time pressure is a fallible component, and infrastructure should reduce the number of near-miss opportunities rather than train people to handle them faster. That is a design stance, not a psychological one, but it follows directly from the evidence: if near-miss feedback shortens decision time without reliably improving accuracy, the safest move is to reduce the frequency of near misses at the geometry level.

The 0.6-second question for Polish roads

Poland's road programme has spent the past decade rebuilding junctions, adding roundabouts, and rolling out intelligent transport systems on corridors like the S7 and the A2. The next layer of investment is likely to be data: connected signals, in-vehicle warnings, and municipal traffic-management centres that can observe conflict points in real time.

That creates a choice. Feedback systems can be tuned to make drivers faster — a tempting metric for throughput — or to make them better calibrated. The 0.6-second result suggests the first is easy and the second requires deliberate design. A junction that tells you, honestly and early, that the gap is closing is doing something different from one that lets you discover it at the last moment.

The forward-looking move for Polish cities is to treat near-miss data as a calibration instrument rather than a scoreboard. Municipal traffic teams already collect conflict data at blackspots; feeding that back into signal timing, warning thresholds, and junction geometry — and measuring whether decision accuracy improves alongside decision speed — would turn a laboratory finding into an engineering standard. The question is not whether feedback shortens decisions. It does. The question is whether we design the feedback to shorten the right ones.