The first ten minutes in a car you do not know. Your driving is fine. It is calibrated to a different car.

Most of what you learned on lessons is portable. Where to look approaching a junction, how to read a gap, when to commit and when to wait, what a wet road does to your margins. None of that lives in the car. It lives in you, and it walks from vehicle to vehicle unharmed.

A second, quieter set of skills does not travel at all. It is a collection of physical settings you learned by repetition in one specific machine:

• where the pedals sit relative to each other, and how high, wide apart and heavy they are
• the point in the brake pedal's travel where the car actually starts slowing
• the clutch weight and the height of the biting point in a manual
• mirror geometry, and therefore the shape and size of the blind spot
• which side the indicator stalk is on, and which stalk does the wipers
• where the demister, the rear fog lights and the headlight dip live
• how wide the car is, where its corners are, how far the boot sticks out behind you
• what the car does on its own when you do nothing at all

Swap car and the first list is intact while the second is wiped. You feel that as clumsiness and read it as a bad day. It is neither. It is a calibration problem: the way a car talks to you through its controls is exactly the channel that goes quiet in an unfamiliar vehicle.

The police manual Roadcraft puts vehicle control at the bottom of a four-level model of driver competence, and says outright that an unfamiliar vehicle raises your risk. It is a UK manual and none of its procedures are Irish law, but the structural point holds: the base layer is car-specific, and you rebuild it every time you change car.

The strongest evidence here is not from driving schools. It is from the US National Transportation Safety Board, which investigated a set of pedal-misapplication crashes and published them as a highway special investigation report.

Reviewing the research on unintended acceleration, the NTSB recorded that such incidents "were also found to occur most frequently in vehicles with which the driver was unfamiliar." Elsewhere in the same report it names "negative transfer from other vehicles" as one of the sources of variability that can put a driver's foot on the wrong pedal, alongside head and body position, gaze direction and optical flow.

Two of its cases are worth knowing, because both drivers were professionals in a substitute vehicle:

• Falls Township, Pennsylvania, 12 January 2007. A school bus driver on an ordinary route, driving a replacement bus. His usual bus was front-engine with a bulkhead-mounted accelerator; the replacement was rear-engine with a floor-mounted accelerator, and the pedals differed in size, height and separation. No mechanical defect was found. The NTSB determined the probable cause was pedal misapplication, and that contributing to it was the driver's unfamiliarity with the school bus.
• Nanuet, New York. A driver on her first scheduled run in a replacement bus whose pedals differed from her usual one. She reported afterwards that the brakes had failed. Examination found no apparent defect.

The Board's finding on the pair is blunt: those two accidents demonstrate that unfamiliarity with the pedal configuration of an alternate bus may lead to error and pedal misapplication. An earlier NHTSA review, summarised in the same report, had already listed driver familiarity with the vehicle among the possible contributing factors to sudden acceleration, alongside pedal placement, feel, travel and offset.

Why it shows up when you are startled, not when you are thinking

The error is not a knowledge error. You know perfectly well this car is an automatic. Under pressure you do not reason, you reach, and the reach is aimed where the control used to be. Learners meet the harmless versions constantly:

• a hand going right for an indicator that is on the left in a Japanese import, so the wipers sweep a dry windscreen at a roundabout
• a left foot searching for a clutch that does not exist in an automatic
• a hand groping at the space between the seats for a lever, in a car whose parking brake is a small switch
• reaching for a headlight dip and finding the rear fog lights

The idea that we revert to earlier-learned routines under stress goes back to a 1959 experiment by Barthol and Ku, in which people were taught two ways of tying knots and then put under pressure. Old, small, and about knots, so take it as the name of a plausible mechanism rather than proof about drivers. The weight here sits with the crash investigation.

The practical consequence is the reason this article exists. Rehearse the differences deliberately, out loud, in the first minutes, while you are stationary and calm enough to be clumsy safely. That is cheaper than finding them in the half second when a child steps between parked cars.

Every distance you judge for the rest of the drive is measured from where your eyes happen to be. Sit six centimetres lower than usual and the bonnet eats road, kerbs move, and your sense of where the front of the car ends quietly goes wrong. First job, before the engine. The order we teach, seated with the door open if you need room:

1. Seat distance. Full clutch or full brake travel with the knee still slightly bent, not locked.
2. Seat height and backrest. As high as headroom allows for a new car, so you can see the near kerb.
3. Head restraint. The top of it level with the top of your head, not resting under your skull.
4. Steering wheel reach and rake. Wrists able to rest on the top of the wheel with shoulders still against the seat.
5. Belt. Flat across the shoulder and low on the hips, not twisted, and it should lock when you tug it.
6. Door properly shut. Cars from other people's driveways get slammed carelessly.
7. All three mirrors last, because moving the seat moves the mirrors' usefulness.

Two Irish details make this more than good manners. First, it is a marked aspect in its own right: the RSA's guidelines put "the door, mirrors, seat, seatbelt, head restraint adjustments" with technical checks under aspect 1, Rules/Checks, where three or more deficient checks or wrong answers earns a fault, to a maximum of one. The tester asks out loud, in a fixed line: "would you please ensure that the seat, seat-belt, head restraint, and mirrors are properly adjusted and that your door is properly closed."

Second, you are allowed help with the mirror you cannot reach. The procedures say the tester should adjust an exterior mirror if asked to by the applicant, and should assist with the left-hand door mirror if requested, readjusting it after the reverse. In an older borrowed car with no electric adjustment on the passenger side, that is the difference between a set mirror and a guess.

Do that arithmetic once and you never do it again. At 60 km/h you cover about 16.7 metres every second. A two-second search for an unfamiliar demister switch is more than 30 metres of road travelled with your eyes inside the car. That is the whole argument for finding switches while the handbrake is on.

In Ireland you do not have to guess which switches matter. The RSA publishes two lists and examines both, at two different moments, and they are not the same list.

Outside the car, before you get in: the technical checks

The Standard Procedures put it like this: the applicant should be asked to verbally explain how they would perform a check on any three of the following aspects, selected at random by the driver tester.

One detail catches people out in a car they have never opened. If engine oil, coolant, or wipers and washer fluid comes up, the procedures state the bonnet will have to be opened and closed again securely by the applicant. The catch sits somewhere different in every car, and some need a hand under the leading edge for a secondary latch. Open it once in the driveway.

Inside the car, engine running, before you move: the secondary controls

This is the core of the problem, written into a national procedure. The tester says: please start up the engine, and before you move off, show me how you would operate some of the secondary controls. Then they pick three of ten.

Note the RSA's order, the reverse of what most people assume: engine on and controls demonstrated first, seat, belt, head restraint, mirrors and door second. We teach it the other way round on lessons, because body position is what every later judgement is measured from. There is no single canonical order, for what it is worth: the police manual starts the engine and reads the instruments before it adjusts the seat too. Either way the point stands. You will be asked to operate three switches in this car before you have driven a metre of it.

The marking arithmetic is worth knowing, because it is kinder than most candidates fear. In the Vehicle Controls reference chart, inability to demonstrate one secondary control, or describe one technical check, sits in the band where no mark is recorded. Three is where it turns: "inability to demonstrate 3 secondary controls" and "inability to describe 3 technical checks" are Grade 2. Three is exactly the number the tester asks. Fumble one and you are fine. Know none and you have manufactured a Grade 2 before moving.

There is a marking consequence on the road too. Under Vehicle Controls the guidelines list "driving while condensation restricts visibility, or not using secondary control as required." Failing to find the demister in a fogged-up borrowed car is markable, not merely awkward. Our page on the tell me and show me questions drills the same ground; the lists above are the RSA's own wording.

The brake is the most consequential unknown in any car you have not driven. Two cars of the same size can want visibly different pedal pressure for the same deceleration, and there are only two ways to find out which sort you are in: deliberately, at low speed, on an empty road, or accidentally, at speed, behind somebody's boot. Do the first one.

The stationary check

• Press the pedal firmly. It should move freely to a firm, positive resistance and hold that pressure for three or four seconds without sinking towards the floor.
• Confirm the parking brake physically holds the car, not just that it clicks.
• Note whether the car will start at all without your foot on the brake. Plenty of newer cars will not, and a driver who has never met that behaviour concludes the car is broken.

The low-speed moving check

On a quiet residential road with nothing behind you, roll at walking pace and make one deliberate, progressive squeeze to a stop. Hold the wheel lightly. Feel for the car pulling to one side, for vibration or pulsing through the pedal, and listen for grinding or binding. Then repeat slightly firmer and note how much pedal this car needed.

In a manual, do the same for the clutch: find where it takes up, on the flat, before you need it on a hill with a car behind you. Then carry the finding into your following distance. The RSA marks position following as its own aspect, expecting a safe distance related to the speed at the time, and an unknown bite point plus a tight gap is how drivers find both at once. Our guide to stopping distances covers those gaps.

Ten years ago you could describe both of these in one sentence each. Now they are the parts of a car most likely to be somewhere you have never seen.

Selectors. A manual gate, a P R N D lever, a rotary dial, a column stalk, or a row of buttons. In an electric car, lifting off the accelerator may slow you noticeably through regenerative braking, which changes how much brake pedal you need and is worth feeling once before you rely on it.

Parking brakes. A lever between the seats, a foot pedal, or a switch you pull. Some cars add auto hold, which keeps the brakes on at a stop until you press the accelerator. Some add hill-start assist, which holds brake pressure for a moment after you lift your foot and feels exactly like a stuck brake if you have never met it.

Say it out loud before you move: this is a rotary selector, this is an electronic parking brake, this car holds itself at lights. Naming a control is how you stop yourself reaching for the last car's version of it.

The Irish rules here are specific and quotable:

• The RSA accepts either type. Under Vehicle Controls the marking guidelines note that on a hill start the vehicle must be secured with its parking brake, "i.e. manual or electronic brake whichever is fitted." An electronic switch is no disadvantage on test.
• But it has to work. If the parking brake is not applied or not properly applied, the tester asks the applicant to secure the vehicle; if it is seriously defective the test is deemed non-conducted, the tester returns to the centre and records an N4 code.
• The test is not over until the car is parked properly. The tester must remain in the vehicle until it has been parked, the parking brake fully applied and the engine switched off. "Not applying the handbrake at the end of the test" is listed as a Vehicle Controls fault.
• Manual driver in an automatic: your left foot now has no job and must not drift. The guidelines require the right foot for braking in normal driving, allowing the left only for slow manoeuvring in confined areas such as a turnabout, reverse or park, if done competently. A candidate who insists on left-foot braking after being advised gets a return to the centre and a non-conducted report.

Two small mercies for anyone learning the width of a new car. The guidelines say that during the turnabout, a fault should not be recorded where an applicant turns the wheel while the vehicle is stationary, so-called dry steering. That exemption is written into the turnabout section and nowhere else, so do not stretch it into a general permission. And a kerbed wheel on a left turn is recorded under position turning left rather than steering.

If your lessons have been in an automatic and the family car is a manual, or the reverse, practise it rather than hope. Our page on automatic driving lessons sets out how the two feed into each other, including the licence consequence below.

Assistance is a property of this specific car, not of "modern cars" in general, and the gap between those two ideas is where risk compensation lives. Assume the car will catch your mistakes when it will not, and you have given away margin.

Registration date is a rough sorting tool. Regulation (EU) 2019/2144, generally called GSR2, applied to all newly registered M1 cars from 7 July 2024, bringing intelligent speed assistance, advanced emergency braking, drowsiness and attention warning, and emergency lane keeping. Those dates come from Thatcham Research and RAC explainers rather than the Regulation itself, so check the car's handbook. A 2019 family hatchback probably has none of it; a car registered from mid-2024 has all of it.

Then the limits, which Roadcraft sets out at length while framing assistance as a co-pilot rather than a replacement:

• Blind spot monitoring should not replace mirrors or a shoulder check. It is not designed to pick up parked cars, pedestrians, animals or street furniture, and it may fail in bad weather.
• Autonomous emergency braking may only operate at low speed on a given car, which is a handbook question, not a guess.
• Lane support depends on visible lane markings and fails where lines are faded, obstructed or absent, which describes a lot of Irish rural road.
• Automatic parking struggles in poorly marked bays, on uneven ground and in bad weather.
• Hill-start assist holds brake pressure for a few moments after you lift off, which is the single behaviour most likely to be mistaken for a fault.

The same manual gives the counterweight to any "modern cars are safer" framing: safety technology has moved fast enough to give drivers a false sense of security, and a car with far more safety features than the one you learned in can lead you to take risks you would not have taken before. A newer borrowed car is not automatically the safer choice for an inexperienced driver.

For the test, one rule settles it: the RSA's marking guidelines state that the use of Park Assist is not permitted. Neither the guidelines nor the Standard Procedures mention reversing cameras or parking sensors at all, so we will not tell you they are approved. What the guidelines do require is proper observation throughout the reverse, and a screen is no substitute for that. The tester does not request a reverse parking manoeuvre, although a candidate who attempts one will not be prevented.

For what happens to your own skills when the car takes over parts of the job, see driver assistance and skill fade, which covers the switch-off behaviour and the Irish law on distraction warning systems.

Everything above is a skills problem. What follows is a documents problem, and it is the one that sends people home from a test centre without driving.

There is no Irish figure for how many test candidates present in a car they have barely driven, and no published Irish crash breakdown by vehicle familiarity. If anyone quotes you one, ask where it came from.

There is no trial showing that a familiarisation routine reduces crashes. Nobody has done it. What evidence exists points the other way about driver training generally: formal training has not been shown to cut crash involvement, with a Cochrane review returning a risk ratio of 0.98. We publish that inconvenient finding in does driver training actually work. So the honest claim above is narrower than a safety claim: it removes identifiable unknowns, and it covers the two control lists the RSA examines.

The pedal-design literature has no consensus, by the NTSB's own statement, and its cases were heavy vehicles. The rental-car comparison people reach for, Tay and Choi on Korean police data from 2010, found rental crashes more often associated with poor driving, inattention and aggressive driving, and involving more males and more drivers under 25, with no differences in vehicle or environmental factors. Unfamiliarity is a candidate explanation there, not a measured effect, and we could only read the abstract.

The RSA does not require familiarity with the test vehicle. We looked. Nothing in the Standard Procedures obliges a candidate to have driven the car before. That is an absence we observed in the document, not a rule, so do not read it as an RSA endorsement of turning up in a car you have never driven.

And a note on sourcing. The procedures document we read is headed 20 February 2026, while rsa.ie appears to be serving a later edition dated 23 June 2026 at the same address. Procedural detail moves, particularly the interim insurance wording. Check it against the current PDF, and if a tester says something different on the day, the tester is the authority.

1. Get built into the car. Seat distance, seat height, head restraint, wheel reach and rake, belt flat and locking, door properly shut.
2. All three mirrors, after the seat, not before. Ask for help with the far door mirror if you cannot reach it.
3. Walk the ten technical checks in your head and find the bonnet release. Open the bonnet once and shut it securely.
4. Ignition on, engine off. Read the warning lights that come up.
5. Start the engine. Read them again. Anything still lit is a question to answer now, not on the road.
6. Find and work the switches: wipers, washer, dipped and main beam, rear fog, front and rear demist, fan, hazards, horn. Out loud.
7. Name the selector and the parking brake. Which type is each? Does the car hold itself at a stop?
8. Static brake press, firm and held, then confirm the parking brake holds the car.
9. Roll at walking pace and make one deliberate progressive brake application. In a manual, find the clutch bite point on the flat.
10. Say three things this car does that your usual car does not. If you cannot name three, you have not finished looking.
11. Then join traffic, and give yourself more following distance than usual for the first ten minutes.

If the car is a test car, add the documents before any of that: certificate of insurance in your hand, L plates on, tax and NCT discs current, dashcam and interior cameras off, sensible shoes on your feet.