Back to blog

Pilot Written Exam: Card the Condition, Not Just the Limit

Regulations are if-then rules. A card holding only the limit throws away the condition the exam question is built on. Here is the shape that keeps both.

How to study for the pilot written exam, compressed into one line: work a question bank, and every time you miss a question, write a card that starts with the situation instead of the number. The regulations you are being tested on are conditional rules, and the exam almost never asks you what a limit is. It hands you a scenario and asks which limit applies to it.

My name is Daniel. I am a third year engineering science student at UofT, and I am not a pilot, so I am not going to invent a story about a checkride. What I do is build study tools, and I work on Nebulearn, which is a conflict of interest you should know up front. What I can help with is card structure, because conditional rules are what flashcards are worst at by default, and the same failure shows up in law, nursing and accounting decks that look nothing alike.

One habit I am keeping the whole way down: no altitudes, distances, currency intervals or regulation numbers. Aviation regulations differ by country and they get amended, so the only version that matters is the one your own authority publishes today. Every example below uses placeholders like LIMIT and CATEGORY X, and you fill those in from your current regulations document.

A regulation is an if-then, and the limit is only half of it

Strip any regulation down to its bones and you get one shape. If these conditions hold, then this requirement applies. Conditions in the first half, requirement in the second. Nearly every line in the document you are studying is some version of that, sometimes with the conditions buried in a definition three sections away.

Now look at how the exam uses that shape. A question stem is almost entirely first half. It gives you an operation type, an aircraft category, a set of circumstances, and asks for the second half. The difficulty lives in reading the conditions correctly and picking the branch they select.

So a card whose front reads "what is the minimum LIMIT" and whose back is a value has carded the second half and thrown the first half away. You will get good at producing that value. You will not get better at the thing being measured, which is deciding whether this scenario is the one where that value applies.

The mismatch is in the retrieval cue. Your card's cue is the name of the limit; the exam's is a paragraph of situation with no rule name in it. You practised one road and got tested on the other, which is the context dependency problem in its purest form.

What the naked number card teaches you instead

It is worth being specific, because this failure does not feel like failure while you review.

The first way is a false positive. Regulations come in siblings: the same requirement carrying different values for different categories, operations or circumstances. If your deck holds one bare value, you own exactly one answer and you will hand it confidently to the scenario where it is wrong. Nothing in your review history warned you, because the card could not expose that gap.

The second way is knowledge you cannot trigger. You can recite the value on demand and still stall on the question, because the stem never says the name of the rule. You are waiting for a cue the exam does not give.

The third way shows up once you notice the sibling problem and fix it badly, by adding each sibling as its own bare-value card. Now you have four cards whose fronts are near-identical strings differing by one word and whose backs differ by one figure. That is a pair you get wrong forever, and it is usually where someone decides flashcards do not work for regulations. Fronts that do not discriminate are the problem, which is the same trap as making cards that are too easy wearing a different hat.

The card shape that carries the condition and the limit

The fix is a three-line card, and the third line is the one people skip.

Front: the situation, written roughly the way an exam would hand it to you, with every condition that selects the answer stated out loud and nothing that does not. Then the question.

Back, line one: the requirement or limit. Line two: the condition that selected it, named. Line three: the reference, meaning the document and the section identifier as they appear in your copy today.

Line two is the self-grading line: when you get a card right, you check that you got it right for the reason and not by the shape of the card. It is the only real defence against "felt like I knew it" once the deck is a few hundred cards deep.

What you are cardingLoses the conditionKeeps it
A minimum that varies by aircraft category"What is the minimum LIMIT?""Operation type OP, category X, conditions A and B hold. What is the minimum LIMIT?"
A currency or recency requirement"How often must REQUIREMENT be done?""A pilot wants to do ACTIVITY. PREREQ was last completed at time T. What has to be current, and by when?"
An equipment rule"List the required equipment for OPS""An aircraft is operating under OPS in condition C. Which item in this list is not required?"

Read down the right column and notice what changed. The front got longer and the back got shorter. That is the trade you want, and the opposite of most people's instinct to keep fronts short so reviews go fast.

One discipline on the front: add a condition only if it changes the answer. Padding the scenario with circumstances that have no bearing on which branch applies quietly teaches you that those circumstances matter, and the exam punishes that in the other direction. One decision per card, which is atomic encoding applied to conditions rather than to facts.

Cloze works here too, arguably better, because you cannot delete a value out of a full sentence of the rule without the conditions coming along for free. Write the rule as one sentence and hide the limit, something like the minimum is {{c1::LIMIT}}. Then write a second cloze hiding the qualifying condition instead, so you can run the rule both directions. Cloze versus question and answer covers when each earns its place.

Rules that branch, and how many cards they are worth

Single rules are the easy case. Most of the pain in a regulations deck comes from rules that fork, where one requirement has three or four values depending on some dimension, and the exam is built out of that fork.

Three card types handle a branching rule, and they are not interchangeable.

The branch map comes first. Front: "requirement R changes along which dimension?" Back: the names of the branches only, no values. It is the cheapest of the three and does the most work, because it stops you answering out of the one branch you happen to remember. Name the fork and you know to go looking.

Then one card per branch, each front carrying the condition that selects it, built like the right column above.

Then the contrast card, but only for the two branches you keep swapping. Front: "same scenario as before, except CONDITION flips from A to B. Does LIMIT change, and in which direction?" A directional answer is often what the question is really testing, and "it goes up" is much cheaper to hold than two absolute values you are already blending together.

What to avoid is the table card, the whole four by three matrix pasted onto one back. You will read it, nod, and hit good. Reading is not retrieval. Keep the matrix in your notes as the thing you build cards from.

On count, be stingy. A rule with three conditions and two outcomes each is eight cards if you are a completionist, and you would have gotten most of those right without making them. Build the branch map, then card the branches you have actually missed. An exam deck audit a few weeks in tells you which those are.

Where the question bank fits, and the two-question triage

Everything above is a shape. The question bank is where you find out which shapes you need, and it turns "study the regulations" into a week you can schedule.

Work a block of bank questions first, then go back through the misses and ask one thing about each: did I not know the limit, or did I not see the condition?

If you did not know the limit, that is a conditioned limit card. Straightforward.

If you knew the limit and applied it to the wrong scenario, do not make the limit card. That miss needs a contrast card or a branch map; adding the value again only makes the review feel productive.

If you read the stem and did not notice a condition was present at all, that is a third kind of card. It is rare, and it is the one that fixes scenario blindness. Front: the stem sentence. Back: which words in it changed which rule applies.

One rule about sourcing. Card from the regulations document, not from the bank's explanation blurb. The explanation is a third party's paraphrase; the exam is written from the source. Use the explanation to find the section, then card the section, keeping the wording close. Same evidence-then-claim discipline as pulling cards out of a PDF, and the reason the reference line exists. For pacing across the weeks you have, the high stakes exam schedule post is what I would read next.

When this method breaks

It does break, in four predictable ways, and three of them have fixes.

Your document gets amended, and cards go stale silently, which is the worst kind of wrong. This is what the reference line is for: when your authority publishes a change, search your deck for the section identifiers it touched, reread those sections, and rewrite rather than trusting the card. Cards with no reference line cannot be audited that way, so rebuild them from the source.

The deck outgrows the week. Skip or ignore the cards for branches you do not need yet instead of deleting them. A rule you will not study until next month still should not be in tomorrow's queue.

The same pair keeps coming back wrong. Five failures on one pair is not a repetition problem, it is a card problem. Collapse the pair into one contrast card and let the absolute values follow later.

And some of this exam is not cards at all. Chart reading, weight and balance, performance interpolation, plotting: those are procedures, learned by doing problems. Cards hold the base layer so you remember that a rule exists and when it fires. Problems teach you to execute. I run the same split for circuits.

What the split means for the tool you use

The argument on this page is one split: condition versus limit, kept on the same card. What that costs you in practice is friction, because you are reading a long regulations PDF in one place, a bank explanation in a browser tab somewhere else, and writing cards in a third window while trying not to lose the section number.

Here is the honest part where other apps win. If you study somewhere with no signal, Anki is local-first and wins outright: Nebulearn has no offline mode, and pretending otherwise would be a lie you would discover on a bad day. If your ground school already has a big shared deck going, Anki wins that too, though your review history will not follow you anywhere you later take the deck. And for a Friday evening quiz session with two classmates on one shared set, Quizlet is still the path of least resistance.

For the job this page is about, getting conditions and limits out of a long document and into cards that keep their reference, PDF upload with a side-by-side view is the feature that matters, and it is what I would point you at in Nebulearn: the source stays open beside the card you are writing, so the condition and the section identifier come from the document rather than from memory. FSRS is on by default, which is what makes branching rules survivable, because the branch you keep missing comes back soon and the branch you have cold waits. The free tier holds 1,000 saved questions and 40 weekly AI credits, more than one regulations subject needs.

pilot-examflashcardsstudy-tipsspaced-repetition