Exam strategy

What you are sitting

The Foundation Level exam is 40 questions in 60 minutes, and the pass mark is 65% — 26 of 40. If you are not sitting it in your native language you are entitled to 15 extra minutes, which is offered as a checkbox when you start our practice paper.

Our practice paper is built to the published exam structure, and you can plan around the shape of it. The question count, duration, extra-time allowance, pass mark and the per-chapter split all match the structure ISTQB publishes for CTFL v4.0:

Ch	Topic	Questions	K1	K2	K3
1	Fundamentals of Testing	8	2	6	0
2	Testing Throughout the SDLC	6	2	4	0
3	Static Testing	4	2	2	0
4	Test Analysis and Design	11	0	6	5
5	Managing the Test Activities	9	1	5	3
6	Test Tools	2	1	1	0
	Total	40	8	24	8

Read that table once before you revise, because it prices your effort. One honesty about it: our draw matches the question counts, not the K-level columns — we draw the right number of questions per chapter, and we do not yet guarantee that exactly five of chapter 4's eleven are K3. Use the K columns to plan your revision, and the chapter counts to predict our paper.

The timing arithmetic, and why uniform pacing is wrong

Sixty minutes over forty questions is 90 seconds each. Nobody should actually pace that way, because the questions are not the same size:

Question type	Realistic time	Why
K1 recall	20–30 seconds	You know it or you do not; staring does not help
K2 explain or compare	60–90 seconds	Read the stem carefully, eliminate, decide
K3 apply	2–3 minutes	You have to actually work something out

So the plan is: bank time on the recall questions and spend it on the K3 ones. A first pass answering everything you know quickly should leave you fifteen to twenty minutes for the handful that need real work.

Where the slow questions live is knowable in advance, and the table above prices it exactly. Every K3 objective in the syllabus sits in chapter 4 (the four black-box techniques and ATDD) and chapter 5 (estimation, prioritisation, defect reports) — and the published structure agrees: all eight K3 questions are in those two chapters, five and three, and every other chapter is K3-free. Those two chapters are also 20 of the 40 questions, so half the paper is the two chapters that hold all the calculation, and the other half is almost entirely recognition.

Put the two facts together and the arithmetic closes: eight K3 questions at two to three minutes is roughly twenty minutes, which leaves forty for the other thirty-two. That is the whole pacing plan — the remaining thirty-two average 75 seconds, and the twenty-four K2 questions are what actually consume it, since eight K1 questions cost about four minutes between them.

The question styles, and how each one breaks

Single best answer. Four options, one correct. Eliminate two quickly, then decide between the survivors. If both survivors look right, the stem contains a qualifier you skimmed.

Multiple response — "select two" or "select all that apply". In our practice paper these are graded as an exact set: every correct option and no incorrect ones, with no partial credit. Selecting one of two correct answers scores the same as selecting nothing. So if the stem says select two, select exactly two.

Scenario questions. A paragraph of situation, then a question. Read the final sentence first, then the scenario — you will read it looking for something rather than trying to remember all of it.

Negative questions — "which is NOT", "which is least likely". These catch people who read fast and answer the positive version. When you see a negative stem, mark it in your head and check your answer against the stem once more before moving on.

"Best" or "most" questions. More than one option is defensible; you are being asked to rank. These are usually K2 questions about a distinction — the "best" answer is normally the one that names the actual mechanism rather than a true but generic statement.

Words that decide answers

In the options, absolutes are usually wrong:

always · never · only · all · must · guarantees · eliminates · proves

Testing is a discipline built on "it depends" and "reduces the probability of", so an option promising certainty is usually the distractor. Compare:

• "100% branch coverage guarantees 100% statement coverage" — true, and one of the very few guarantees in the syllabus.
• "Passing tests prove the software has no defects" — false, and it is principle 1 rewritten as a trap.

The reverse is also worth knowing: hedged options survive more often. "Usually", "typically", "can", "may" describe how the syllabus actually talks.

In the room

First pass — answer what you know. Do not flag things because they feel hard; flag things where you have narrowed to two options and need another minute. A flagged question you have not answered at all is a question you might run out of time to return to, so always leave an answer behind, even a guess.

There is no negative marking in our practice paper — a wrong answer and a blank answer both score zero. Guessing is strictly better than leaving something empty. With four options and two eliminated, a guess is a coin flip, and a coin flip is worth half a mark on average.

Second pass — the flagged ones, in the order the navigator shows them.

The last ten minutes, and in this order:

1. Every unanswered question gets an answer. This is not optional and it comes first, because it is the only part of the last ten minutes that can still gain you a mark from nothing.
2. Re-read the negative-stem questions you remember flagging. This is where careless marks are recovered.
3. Leave everything else alone. Wholesale second-guessing loses more marks than it gains; change an answer only when you can say why the first one was wrong — a rule you misremembered, a qualifier you missed — not because it feels uneasy.

Watch the two warnings the runner gives you at ten minutes and at two minutes. The paper auto-submits at zero, so anything unanswered at that point stays unanswered.

Preparing, in the week before

1. Drill chapter by chapter. Take each chapter quiz until you are consistently at 6 of 8 or better. Chapters 4 and 5 are the ones to over-invest in — together they are half the paper.
2. Then sit a full timed paper, under real conditions: one sitting, no notes, the clock running.
3. Read the per-chapter breakdown on the result page, which exists precisely for this. A 60% overall that hides 2 of 11 on chapter 4 is a different problem from one spread evenly, and it tells you exactly what to revise.
4. Re-read the explanations for every question you missed, including the ones you guessed correctly — a lucky guess is an unlearned objective wearing a correct answer.
5. Sit a second paper. A different seed draws a different set, so the score means something.

Take the full practice exam →

The evening before, and the morning of

Nothing new. Re-read your own notes on the distinctions — the tables at the end of each chapter lesson are built for exactly this pass — and sleep. A K3 question answered by a tired brain is the most expensive mistake available, because those are the questions with a single unambiguous right answer.

Bring photo ID if you are sitting a proctored exam, and check the platform's requirements the day before rather than the hour before.

You have finished the track

Five tracks: what QA does, professional manual testing, automation, the senior material, and this. Whatever the certificate ends up saying, the useful part was never the certificate — it was learning to say precisely what you tested, what you did not, and why.

Good luck. Then go and test something real.