Grade 2 or Grade 3? How the tester actually decides.
Two learners make the same mistake. One collects a Grade 2 and passes; the other gets a Grade 3 and fails. That is not luck. It is a risk decision with a logic you can learn, and this is how it gets made.
In this guide
This article is adapted from Marking Faults: How RSA Driving Testers Think, the manual we wrote to train our own instructors and mock testers. Here we are sharing the part every learner deserves to understand before test day: the reasoning behind the grade. For the grade definitions and the four fail rules themselves, see our companion guide, Why did I fail my driving test?
The one question running through the tester's head
Risk, not mistakes.
Most learners imagine the tester as a referee with a whistle, waiting for infringements. The picture we teach our instructors is different: think of the tester as a risk auditor in the passenger seat for half an hour. Every action you take passes through one filter: did that create, increase, or fail to manage risk for another road user?
That single question explains most of what confuses people about results. Learners think in checklists: mirror, signal, manoeuvre. The tester is watching outcomes: did anyone have to brake, slow, swerve or hesitate because of you? In our experience, that assessment runs as a handful of parallel checks, fast and almost subconscious:
- Worst plausible outcome. If nothing changes in the next two seconds, what happens?
- Proportionality. Was your action appropriate to the hazard in front of you?
- Prior knowledge. Did you know the hazard was there before you acted, or did you get lucky?
- Third-party effect. Did another road user have to change what they were doing?
When the answer to that last check is "yes, significantly", the fault is heading for Grade 3. When it is "yes, mildly", Grade 2. When it is "no, but the technique was weak", Grade 2 is still likely, because in practice testers no longer record minor faults: anything worth writing down at all now carries weight.
One thing this is not: arbitrary. Two testers may differ slightly on the small stuff, because they see head checks from different angles and may draw the pattern line in slightly different places. What they do not differ on is actual danger. Force another driver to brake hard and every tester in the country records the same grade.
What the RSA itself puts in writing
The official layer, with page numbers.
Before we get to how instructors read marking in practice, it is worth being clear about what the RSA actually publishes, because the two layers are different and both matter.
The RSA's marking guidelines, Making Your Mark, define the grades in one line each: a Grade 2 is "a serious fault", and a Grade 3 is a dangerous or potentially dangerous fault, or total disregard of traffic controls (p. 4). The same page confirms the point above about minor slips: faults falling in the 0% to 35% band of the marking scale are not recorded at all, because they are not deemed serious.
The document's reference chart (p. 24) then puts every marking aspect on a severity scale with three bands: no mark recorded up to 35%, Grade 2 labelled "More Serious" from 36% to 75%, and Grade 3 labelled "Dangerous or Potentially Dangerous" from 76% to 100%. For each heading it gives worded examples of the bands that apply. For observation, for instance, a slight infringement gets no mark, "not taking proper observation" is Grade 2, and a dangerous or potentially dangerous lack of observation is Grade 3.
Two more official details are worth knowing:
- Faults are not double marked. With certain exceptions, where one action suggests two or more faults, the tester records the one that conveys "the most pertinent information" to you (p. 3). Hit a kerb turning left and the mark typically goes under position turning left, not steering. The sheet is written to teach you, not to pile on.
- The guidelines are examples, not an algorithm. The RSA states plainly that the guidelines "do not purport to cover every possible situation" (p. 3). Judgement is built into the system by design.
And Grade 3 decisions are not casual: the RSA's standard procedures require testers to note the circumstances behind every Grade 3 they record, what happened and where on the route it happened (p. 20 of the 2025 procedures). A Grade 3 is a documented judgement, not a mood.
So the official picture is: clear definitions, banded severity, worked examples, and an open acknowledgement that the examples cannot cover everything. The gap between those examples is where tester judgement operates, and it is what the rest of this article is about. What follows is not RSA text: it is our professional framework, built from instructor experience of real tests, mock tests and result sheets.
The five escalators: what turns a 2 into a 3
Our framework for how faults grow.
Most Grade 3s do not arrive as Grade 3s. They start life as weak technique, the raw material of a Grade 2, and escalate when the situation adds an aggravating factor. Technique decides the starting grade; consequence decides where it ends up. In our manual we teach five escalators, and once you know them, most grading decisions become predictable.
A worked example. Here is the pair of scenarios we use in instructor training, because the technique in both is identical:
| Scenario | What happens | Grade |
|---|---|---|
| Moving out, car behind | Suburban road. You signal and move out to pass a parked car; your mirror check is late, but the car behind is far enough back and is not forced to brake. Weak technique, mild effect, recovered safely. | Grade 2 |
| Moving out, cyclist alongside | Same road, same action, same late check. This time a cyclist is sitting off your right rear quarter. You move out without an effective check and the cyclist has to brake hard. | Grade 3 |
Nothing about your driving changed between those two rows. The consequence changed, and the grade followed it. That is the core of tester thinking, and it is why memorising "moves" misses the point: the same move is safe in one context and dangerous in another. Judging the context is the skill being tested.
What looks worse than it is, and what looks better
Drama and risk are not the same thing.
Because the grading follows risk rather than embarrassment, the loud moments of a test and the dangerous moments of a test are often different moments. Learners consistently misjudge both directions.
Looks worse than it is. These feel like disasters in the driver's seat but usually stay at Grade 2:
- A loud stall on a quiet road, recovered calmly with nobody close behind. Embarrassing, noisy, and low-risk.
- A wide but safe turn into an empty side road. Poor positioning, no one affected.
- A slightly rough gear change or a moment of clunky steering. Smoothness is not what is being graded; risk is. Even crossing your hands on the wheel is not, by itself, on the fault list, as we cover in our guide to steering and the driving test.
Looks better than it is. These feel smooth and confident from the inside and are the ones that quietly fail tests:
- A calm, undramatic move-off where you never looked over your right shoulder, and a cyclist had to slow. No noise, real risk.
- A confident roundabout entry that was actually a poor gap decision. The car on your right eased off, nobody beeped, and it is still a Grade 3, because they should never have had to adjust.
- A controlled pass of a parked car that drifted across the centre line into the path of oncoming traffic that happened to be slow-moving. Composure does not neutralise position.
This is the hardest lesson in test preparation: a clean-looking drive is not always a safe drive, and a scrappy-looking drive is not always a failing one. It is why candidates walk out saying "it felt fine" holding a fail, and why others apologise their way through a stall and pass. If that has happened to you, our free result analyser will translate the sheet into what the tester actually saw.
How to actually use this
Drive to manage risk, not to avoid mistakes.
Knowing the escalators changes how you prepare. The goal is not a mistake-free drive; nobody manages that for forty minutes, and the tester knows it. The goal is a drive where your mistakes never acquire consequences, and that is trainable:
- Ask the tester's question yourself. After every junction, lane change and move-off: did anyone have to do anything because of me? If the honest answer is ever yes, that is the habit to fix, whatever your instructor's checklist says about the rest.
- Protect the vulnerable first. Escalator 2 means cyclists and pedestrians deserve your widest margins and your earliest decisions. Space and reduced speed around them buy back every other weakness.
- Be predictable. Signal early, brake progressively, position deliberately. A driver who telegraphs every intention rarely surprises anyone, and surprise is what escalates faults.
- Treat legal lines as absolute. Stop means a full stop. A red means stopped. These are the cheapest Grade 3s to avoid because they involve no judgement at all.
- Hunt your patterns before the tester does. One slip is a lapse; the same slip repeated is evidence. This is where an outside eye matters, because you cannot see your own habits from the driver's seat.
These habits are easiest to build early, which is why we bake the "effect on others" question into the very first lesson. In the final weeks, the most valuable exercise is a mock test marked the way this article describes: same sheet, same escalation logic, debriefed grade by grade. Pair it with our test-day guide so the admin side holds no surprises either.
Quick answers
The questions learners ask about grading.
How does the tester decide whether a fault is Grade 2 or Grade 3?
The RSA defines Grade 2 as a serious fault and Grade 3 as a dangerous or potentially dangerous fault, or total disregard of traffic controls. In practice the deciding question is consequence: a minor effect, recovered safely, is Grade 2. If another road user was forced to brake, swerve or stop, or you crossed a clear legal line like a stop line or red light, it becomes Grade 3 and the test is failed.
Can the same mistake be Grade 2 on one test and Grade 3 on another?
Yes, and it is not inconsistency. Identical technique can produce different grades because the surrounding situation was different. A late mirror check with nobody nearby is a Grade 2; the same late check with a cyclist alongside who has to brake is a Grade 3. The tester grades what your action did to the situation, not the action in isolation.
Do driving testers record every mistake they see?
No. The RSA marking guidelines state that faults in the lowest band of the marking scale, 0 to 35 percent, are not recorded because they are not deemed serious. Faults are also not double marked: where one action suggests two or more faults, the tester usually records the one that conveys the most pertinent information to you. What testers do consistently record is anything serious, and repeated versions of the same weak habit.
Is stalling a Grade 2 or a Grade 3 fault?
It depends on the consequence, which is exactly how escalation works. A single stall on a quiet road, recovered calmly with nobody close behind, is typically a Grade 2 at most. A stall on a hill that lets the car roll back towards a vehicle close behind, or repeated stalls that build into a pattern, is a different story and can fail the test.
How many Grade 2 faults fail the driving test in Ireland?
You fail with 9 or more Grade 2 faults overall, 6 or more under one heading, 4 of the same Grade 2 fault on a single aspect, or a single Grade 3 fault. Our guide to why people fail the driving test covers the four fail rules and each grade in full.
Are the Five Escalators an official RSA system?
No. The Five Escalators are Smart Driving Academy's professional framework, taken from our own instructor training manual. The RSA publishes the fault definitions, grading bands and worked examples, and its guidelines state openly that they do not purport to cover every possible situation. The escalators are how experienced instructors read the space between those examples.
See your drive the way a tester sees it.
Our mock test is marked to the same sheet with the same escalation logic described here, then debriefed grade by grade. Already have a result sheet? The free analyser explains every mark on it.
Sources
- RSA: Making Your Mark, Marking Guidelines for the Driving Test (grade definitions and the unrecorded 0 to 35% band, p. 4; no double marking and scope of the guidelines, p. 3; reference chart severity bands, p. 24)
- RSA: Standard Procedures for Conducting Driving Tests (2025) (recording the circumstances of Grade 3 faults, p. 20)
- Smart Driving Academy: Marking Faults: How RSA Driving Testers Think (internal instructor training manual, updated edition 2026). The risk-based reading, the Five Escalators and the worked examples in this article are our professional framework from that manual, not RSA-published text.