This post tests the answer tips that students share before every exam against every A/L ICT Paper I answer key from 2011 to 2026. Three tips are the most common: "Option 1 is rarely the answer." "Three of the same answer in a row means one is wrong." "If you must guess, pick 3." None of them holds. The keys behave like random keys, so a real pattern and a lucky streak can be told apart only with numbers, and Paper I marks come from knowledge and elimination.
The post has four parts:
- The data: the 800 keys, where each one comes from and which ones are left out (section 1).
- The tests: option counts, runs and repeats, and question numbers (sections 2 to 4).
- The guessing tricks: what four tricks would score on 10 guessed MCQs (section 5).
- What works instead: elimination, the marking rules and past papers to try (section 6).
1. The 800 keys we tested
The tests use 800 keys from 16 papers, and 783 of them have one correct option. A tip is only as good as the data behind it, so the data comes first.
Paper I has 50 MCQs with five options each. We took the correct option of every MCQ in 16 papers, from 2011 to 2026. That is 800 keys.
- Official keys: the Department of Examinations marking schemes for 2018, 2019, 2020, 2021 (held 2022) and 2025.
- Copies of the scheme's answer page: 2011 to 2014, 2016, 2017, 2022 (held 2023), 2023 (held 2024) and 2024. These copies circulate on past paper sites, not from the Department.
- Our own keys: 2015 (no copy of its Paper I answer sheet was found) and 2026 (no scheme yet). These are our own worked answers, not official.
Some questions accept more than one option, or every option. For example, 2018 P-I Q40 accepts all five: its key is printed as "All" (2018 marking scheme, page 12). The Paper I answer sheet is one sheet for all three media, with Sinhala, Tamil and English headings ("ලකුණු දීමේ පටිපාටිය", "I පත්රය"), so a key is the same in every medium. The 2021 English-medium and Sinhala-medium schemes carry the same sheet on the same page (2021 marking scheme, English medium, page 16) (2021 ලකුණු දීමේ පටිපාටිය, සිංහල මාධ්ය, පිටුව 16). Our folder has the 2018 scheme in English and Tamil medium only, so the 2018 page has no Sinhala-medium partner. We left those 17 questions out of the tests. That leaves 783 keys with one correct option.
The 2019 old-syllabus paper is also left out. 41 of its 50 keys are the same as the 2019 new-syllabus paper, because the two papers shared questions. Counting both would count the same answers twice.
The grid has no stripe down a column and no block of one colour. Still, most people who look at it for a while see "something". That feeling is the reason this post uses numbers, not eyes.
2. Is every option used equally
No option is the answer more often than chance allows. The most common tip is that one option wins more often. If keys are random, each option should be the answer about one time in five.
Here are the counts from the 783 keys:
- Option 1: 138
- Option 2: 156
- Option 3: 172
- Option 4: 165
- Option 5: 152
An equal share would be about 157 each. Option 3 is a little above, and option 1 is a little below. The question is whether that gap is real.
A chi-square test answers this. It asks how often random keys would give a gap this big. The answer is about 37 times in 100 (p = 0.37). So this gap is ordinary.
Inside one paper, the counts swing much more. 2025 P-I used option 3 for 17 of its 50 questions. 2021 (held 2022) P-I used option 1 only 4 times. 2022 (held 2023) P-I came out almost even. We tested each year on its own. No year is uneven enough to call it a pattern (the lowest is 2011, p = 0.08).
Do examiners balance the options
Examiners do not balance the options. Some students think examiners put each option exactly 10 times. If that were true, every paper would look like 2022. It does not. We compared the spread of the 16 yearly counts with the spread of random keys. The real keys are not more even than random keys (p = 0.72). There is no sign of balancing.
3. Do answers repeat or run
Answers repeat and run at the normal rate for random keys. The second tip is about streaks: "the same answer twice in a row is rare, and three in a row is a mistake".
There are 751 pairs of next-door questions with single keys. In 145 pairs, both questions have the same answer. That is 19.3%. Random keys give 20%. So repeats happen at the normal rate.
Runs are normal too. 10 of the 16 papers have three or more of the same answer in a row. The longest run is five: 2011 P-I Q17 to 2011 P-I Q21 all have option 3. Random keys of 50 questions give runs like this quite often.
We also tested whether one answer predicts the next one, for example whether option 4 often follows option 3. Across the 16 papers it does not (p = 0.20).
4. Does question n keep its answer each year
No question number keeps its answer across the years. The third tip is about position: "Q1 is usually 4" or "the last question is never 5". If examiners had habits, the same question number would carry the same answer in many years.
We counted the keys at each question number across the 16 years. The link between question number and option is not stronger than chance (p = 0.27). Two papers agree on 20.3% of question numbers. Random keys agree on 20%.
The most interesting test is a prediction. For each question, we guessed the option that was most common at that number in the other years.
Looking back, the guess is right 24.2% of the time. That looks like a win. But a real pattern must also predict the future. When we learn from 2011 to 2018 and predict 2019 to 2026, the guess falls to 22.1% (not significant, p = 0.09). When we learn from 2011 to 2022 and predict 2023 to 2026, it falls to 17.9%. That is worse than a blind guess.
Why one test can look lucky
We ran 12 main tests. Even with purely random keys, about one test in 20 gives p below 0.05. So among 12 tests, one "hit" is expected. After we correct for running 12 tests, no result counts as a pattern. A few smaller checks on part of the years gave odd results too. They point in opposite directions in the old and new syllabus years. That is what chance looks like.
5. A thought experiment with ten guesses
No guessing trick beats a blind guess by a margin a student could notice. The thought experiment: a student knows 40 answers and must guess 10, and a trick chooses each guess.
We tested four tricks on the real keys of every year. Each trick used only the other years, or the other answers in the same paper, to choose its guess.
- Always pick the option used most in all other years.
- Pick the option used most at that question number in other years.
- Never pick the same option as the previous question.
- Pick the option used least in this paper. (This is the best case for "balancing". It assumes the student knows all 49 other answers.)
A blind guess gives 2 marks out of 10 on average. Three tricks give about 2.0. The best one, trick 2, gives 2.4 looking back, but it loses in the forward test. Luck alone moves the result by about 1.3 marks either way. So no trick beats luck by a margin a student could notice.
Now think about what a real pattern would do. If examiners always used each option exactly 10 times, trick 4 would be strong. A student who knew 45 answers could count the missing options and guess well above 20%. Trick 4 gets 20%, because the counts are not fixed. If examiners avoided repeats, trick 3 would gain. It gains nothing, because repeats occur at the normal rate.
So many people still believe in patterns because random keys are full of small surprises. A run of three is normal. So is a year with 17 of one option, or a question number with the same answer in 7 of 16 years. People remember these and forget all the places with no pattern.
6. What to do in the exam
The keys give no shortcut, but real skills do raise the score.
Use elimination. Each wrong option removed with knowledge raises the chance of a right guess. With two wrong options removed, a guess is right one time in three. That is better than any position trick we tested.
Answer every question. The marking scheme gives one mark for each correct answer (1 × 50 = 50 in the new-syllabus papers) (2021 marking scheme, English medium, page 16) (2021 ලකුණු දීමේ පටිපාටිය, සිංහල මාධ්ය, පිටුව 16) (2025 ලකුණු දීමේ පටිපාටිය, පිටුව 31). The answer sheet says "එක් පිළිතුරකට ලකුණු 01 බැගින්" and "1 × 50 = 50". It takes nothing away for a wrong answer: the examiner counts only the correct answers. The Sinhala instructions say "නිවැරදි පිළිතුර ✓ ලකුණකින් ද, වැරදි පිළිතුර 0 ලකුණකින් ද" (2021 marking scheme, English medium, page 2) (2021 ලකුණු දීමේ පටිපාටිය, සිංහල මාධ්ය, පිටුව 2). A blank answer is a sure zero; a guess is right one time in five. Mark one option only: the examiner strikes off a question with more than one option marked, or none. The old-syllabus 2018 paper gave 2 marks for each answer, "Total Marks 50 × 2 = 100", although its answer sheet still prints "01" mark for each answer (2018 marking scheme, pages 1 and 12).
Trust the working. Keep a worked-out answer, even if it makes a streak. The keys do not care about streaks.
Spend the time on content. The data says the answer position is random. Marks come from knowing why the right option is right.
Past papers to try
Try these full papers with the timer on. Use elimination on every question that is not certain.
Past PaperA/L ICT 2026 Past PaperMCQ 50 · මිනි. 120 · ලියූ වාර 3Open paper Past PaperA/L ICT 2025 Past PaperMCQ 50 · මිනි. 120 · ලියූ වාර 0Open paper Past PaperA/L ICT 2024 Past PaperMCQ 50 · මිනි. 120 · ලියූ වාර 1Open paperSources
- G.C.E. (A/L) Information and Communication Technology marking schemes, Paper I answers, Department of Examinations, Sri Lanka: 2018, 2019 (new and old syllabus), 2020, 2021 (held 2022) and 2025.
- Copies of the Paper I answer page of the marking scheme for 2011 to 2014, 2016, 2017, 2022 (held 2023), 2023 (held 2024) and 2024, from pastpapers.wiki.
- G.C.E. (A/L) ICT question papers, Paper I, 2011 to 2026, Department of Examinations, Sri Lanka.
- 2015 and 2026 keys: our own worked answers, not official.
- The Paper I answer sheets and marking instructions cited above: 2018 marking scheme (English medium), 2021 marking scheme (English medium and Sinhala medium) and 2025 marking scheme (Sinhala medium).
- Tests: chi-square tests for the option counts, and shuffle (permutation) tests for runs, repeats, next-answer links and question numbers. Each year's own keys were shuffled 4,000 times, which keeps each year's counts and destroys any order.
