The stop you were never tested on. Braking hard in a modern car.

Open the RSA's Standard Procedures for Conducting Driving Tests, the tester handbook dated 20 February 2026, at the contents page and the exercise is listed, with the answer built into the title:

Directly under that heading, on p.30:

A1, A2 and A are motorcycle categories. That "must" lives entirely inside the motorcycle section, and no equivalent sentence appears anywhere in the category B material, category B being the ordinary car licence nearly every Irish driver holds. Even mopeds are exempt, because the AM test "is carried out over a suitable Cat B route", and a car route has no braking exercise in it. The marking guidelines agree: in Making Your Mark the manoeuvre sits under "Motorcycles", assessing "a motorcyclist's ability to deal with an emergency situation".

The car test's set manoeuvres appear in the RSA's note about skipped ones: "Any non- conducted manoeuvres (Reverse, Turnabout, Hill Start)". Those three, plus a park near the end, inside a minimum 25 minutes on the road. Not one second of it near the limit of the brakes.

The one hard stop the RSA does script

Here is the tester's wording for the motorcycle exercise, worth reading for what it does not ask for:

So Ireland's only assessed emergency braking exercise is a speed reduction, and the marking guidelines say that locking up the rear wheel may have a fault recorded for "Brakes". Control and restraint, then, rather than maximum effort, because a rider who locks a wheel ends up on the road. In a car, the electronics took that decision away from the driver, which is why the car advice is close to the reverse of the bike advice.

The footbrake is assessed on a car test, but only as ordinary driving: a candidate should make proper use of the accelerator, clutch, gears, footbrake, handbrake, steering and secondary controls, or a fault may be recorded for "Vehicle Controls". Then comes the line that makes this article necessary, among the examples of those faults:

Sharp braking is markable. There is a second route to the same place under "React", the heading for reacting promptly and properly to hazards, where the examples include "Where an applicant brakes hard on an amber light when the vehicle should clearly have carried on" and "Where an applicant stops suddenly when e.g. turning left on a slip lane by misreading the main lights".

Keep the wording honest. Those are examples, each carrying "may be recorded", and the guidelines say they do not cover every situation. Nobody is banned from braking hard, and an emergency created by somebody else is treated separately: where a tester takes control because of "an emergency which is not the direct responsibility of the applicant, a fault may not be recorded".

But the direction is unmistakable. A candidate is rewarded for never needing to brake hard, and can be marked down when they do. The nearest thing to a braking assessment happens standing still: candidates must "verbally explain" a check on three items chosen at random, the brakes being item (i).

There is a tension here, and we will state it once without turning it into an accusation. Ireland's own road safety strategy, Our Journey Towards Vision Zero, commits under safe vehicles to "Promote and educate on new technologies in vehicles such as: eCall, Automatic Anti-lock Braking System (ABS), Emergency Stopping, Lane Departure". The State says drivers need educating about anti-lock brakes. The State's driving test never asks a car driver to use them.

The pedal goes down much further than you have ever pushed it. Normal driving uses the top third of the travel. A real stop uses the lot, and the first sensation is falling through the pedal with no wall behind it. Keep going.

Then it fights back. A rapid judder against the ball of your foot, sometimes hard enough to feel like a kick. Roadcraft, the UK police driver's manual, describes exactly this as the anti-lock system changing the pressure in the brake lines, and adds a useful point: what feels like the brakes fading may be the system managing pressure to keep the wheels turning. The pedal going strange is information, not failure.

There is noise, grinding or buzzing from under the front of the car, loud enough to sound expensive. The 1995 US public information campaign run by NHTSA, the manufacturers, the ABS Education Alliance and the insurance industry made a point of it: "The ABS system can make noise and vibrate the brake pedal when it is working." And the steering still answers, which is the whole reason it is fitted.

Underneath, your pedal brakes all four wheels together while the system supervises each wheel separately: it senses one slowing towards a lock, releases that wheel's brake, and re-applies the moment it turns again, many times a second. Roadcraft is clear that the repetition is the definition, and a system intervening once in a stop is not doing the job.

Which leads to the condition drivers get wrong. It only works while you keep the pressure on, because there is nothing to modulate otherwise. It is not a brake, it is a supervisor for your right foot. Roadcraft's own instruction is blunter than most learner material dares to be: on a dry straight road, push the pedal as far and as hard as it goes and hold it. No feathering, no releasing. Police-manual guidance rather than an RSA one, but the physics does not change at the border.

1. Pumping the brakes is advice from a different car

It was once correct. Without anti-lock brakes, standing on the pedal locked the wheels, and a locked wheel slides: no steering, just a skid in whatever direction the car happened to be pointing. So drivers released, let the wheel turn and re-applied, doing by hand what the electronics now do far faster. That advice has been wrong for thirty years:

This one is aimed politely at the parents reading. Pumping survives because it gets handed down from the passenger seat by people who learned in cars that needed it. It was true in your first car. It is not true in theirs.

2. Lifting off when the pedal starts fighting

The same campaign spelled it out: "Drivers should not take their foot off the brake pedal when they hear noise or feel pedal vibration." Why did that need writing? Because the evidence on how hard drivers actually brake is grim. Wood and Zhang analysed 3,011 crash and near-crash events recorded by instrumented cars in the SHRP2 naturalistic driving study:

Eighty percent of those events happened on a dry road. Braking experiments collated in the same report put ordinary drivers at about 0.48 to 0.57 g on dry surfaces against 0.75 to 0.83 g for professionals, while type approval requires a car's brakes to reach at least 6.43 m/s², roughly 0.66 g. And drivers who ended up in a crash reacted 0.487 seconds later, and braked 0.018 g more weakly, than drivers in matched near-crashes who got away with it.

So the dominant failure in a real emergency is not skidding, it is not braking hard enough, and the likeliest thing to make a driver back off is the pedal juddering under their foot: the system working.

3. The system built because of all that

Brake assist has been required on new car types in the EU since 24 November 2009, under Regulation (EC) No 78/2009, and the legal definition admits a great deal: a function that "deduces an emergency braking event from a characteristic of the driver's brake demand" and then either "assists the driver to deliver the maximum achievable braking rate" or is "sufficient to cause full cycling of the Anti-lock Braking System". It reads your foot, not the traffic. The EU mandated a system to compensate for drivers not pressing hard enough, and it works by detecting panic in the pedal. So a decisive stab and a tentative one produce two different stops in the same car.

That is not autonomous emergency braking, which brakes without you, and which became compulsory in pedestrian and cyclist detecting form on all new cars and vans registered in the EU from 7 July 2026. It is not retrospective, so nothing is being added to a car already on the road. We cover it in driver assistance and skill fade.

NHTSA measured straight-line panic stops with four-wheel anti-lock brakes enabled and disabled. Reported as medians in its 2009 evaluation:

The gravel result is not a flaw. A locked wheel sliding over gravel builds the loose material into a wedge in front of the tyre, and ploughing that wedge slows the car faster than a rolling wheel does; NHTSA notes snow behaves similarly. The 1995 campaign said it out loud: "ABS can significantly lengthen stopping distances on loose surfaces such as gravel or soft snow." And more bluntly: "Many drivers think the main purpose of ABS is to reduce stopping distances. This is a serious misconception."

It matters more on an Irish back road than it did on an Ohio test track. The default limit on rural local (L) roads has been 60 km/h since 7 February 2025, and those are the roads where grit, gravel wash and mud sit along the edges after rain, in the Meath and Kildare lanes especially. The system is still the right choice, because a locked-wheel slide gives up steering entirely, but stop thinking of it as a distance-shortening device. Our figures on stopping are in stopping distances.

The stability result from the same work is impressive: across 494 straight-line panic stops with the system enabled, not one car yawed, while with it disabled cars yawed on every surface tested. Then NHTSA adds the caveat that never reaches the marketing:

Those numbers came from people who knew exactly what to do. You are not one of them, and neither is anyone reading this, instructors included. The gap between the track and a wet road out of Rathcoole is the driver.

One: in real crashes the system is not the miracle the track results suggest. Across 1995 to 2007 US crash data, NHTSA found close to zero net effect on fatal crash involvements for cars, a 1 percent reduction that was not statistically significant. Non-fatal involvements did fall significantly, by 6 percent. But fatal run-off-road crashes of cars rose by a statistically significant 9 percent, which NHTSA cannot fully explain, offering among other things that "a portion of the public still doesn't know how to use ABS", partly because it "is activated rather infrequently over the life of a vehicle".

Two: the clearest benefit sits where nobody looks for it. Non-fatal at-fault collisions with other vehicles on wet, snowy or icy roads fell 37 percent, which the report calls where the system is most effective, and fatal collisions with pedestrians, cyclists or animals fell 13 percent. Steering while braking, in the wet, in traffic.

Three: that run-off-road benefit arrives with stability control, not anti-lock brakes alone. With both fitted, NHTSA estimates fatal run-off-road crashes down 30 percent in cars.

Four, our standing position on this site: formal driver training has not been shown to reduce crashes. The Cochrane review of post-licence driver education (Ker and colleagues, CD003734) found no evidence that such education reduces crashes or injuries, a pooled risk ratio of 0.98, and we set that out in what actually works in driver education. Nobody should read this article, or feel one hard stop in a lesson, and conclude they are measurably safer. That includes us.

The claim worth making is narrow, and it is the only one the evidence supports: naming these sensations in advance removes one moment of confusion from the worst two seconds of your driving life. That is all, and it is still worth having.

Not on your own, and not on a public road on your own initiative. As a lesson it looks like this: a quiet industrial estate on a dry morning, nothing behind, brakes and mirrors checked, an instructor who has said exactly what is about to happen, and one firm stop from around 40 to 50 km/h, so the pedal shudders once in your life on your own terms. It is not a test manoeuvre, it never will be, and no instructor can honestly promise it makes you safer. Our own learners get it inside the course, set out in our 24-lesson learner course, alongside the twelve mandatory EDT lessons.

If you already hold a licence, the five-minute version costs nothing. Find the ABS lamp on your dashboard during the bulb check, before the engine starts. Spend a few minutes with the handbook working out whether the car has brake assist and autonomous emergency braking, and what the manufacturer calls them. Then learn the three words, stomp, stay and steer, and leave it there.

The grip budget, which decides everything

Each tyre has a fixed budget of grip, spread over a contact patch about the size of your hand, and braking, steering and accelerating all draw on the same account. Spend more of it stopping and there is less left for turning: brake hard and weight shifts forward, the steering loads up, and on a bend the car becomes less stable. Hence the plain rule, finish your braking before you turn, which we take apart in the physics of braking and cornering.

Roadcraft offers a guiding principle that makes most of this article unnecessary on most days: drive so that you can always stop within the distance you can see to be clear on your own side of the road, and on a lane too narrow to pass on, halve it, because an oncoming car needs the other half to brake in. On a Meath back lane at 60 km/h with hedges to the verge, that does more work than any braking technique. The following-distance version, and why two seconds is a floor rather than a target, is in the two-second rule. Again, police manual, not Irish law.

One last point from the same source deserves the ending. Roadcraft's own warning is that safety systems give some drivers a false sense of security, making them over-dependent and prompting risks they would not otherwise take. Safety features cannot change the laws of physics and do not make a driver more skilled. So any time the anti-lock system cuts in, ask what you could have anticipated, and ease your speed for the rest of the journey. A well-executed emergency stop is not an achievement. It is a diagnosis of the thirty seconds before it.