Weak topics from a marked mock test.

The mock comes back marked: 31 out of 50. The number at the top is what everyone looks at. The topics that actually cost the marks are buried in the question-by-question marks, and a topic with two questions can hide behind one with a single question. This is part 4 of the Field Guide series on exam prep with AI: one job, tested on a real screen, with the catch said plainly.

The job, and what to type

Use only your own marks and the topic of each question. Not the question paper, the questions themselves, the answers, or anyone else's marks. The marks below are a sample from a Data Structures mock, made up for this demo: 10 questions, 5 marks each, each question tagged with its topic. Type your own marks and topic names in their place.

Here are the marks from my marked mock test (Data Structures, 10 questions, 5 marks each). Each question has its topic.
Q1 Arrays 5/5
Q2 Linked lists 2/5
Q3 Stacks 4/5
Q4 Queues 5/5
Q5 Trees 1/5
Q6 Linked lists 3/5
Q7 Hashing 2/5
Q8 Trees 2/5
Q9 Sorting 4/5
Q10 Hashing 3/5

- Add up my marks for each topic and show each as marks out of that topic's total, with a percentage.
- Rank the topics from weakest to strongest by percentage.
- Tell me my 3 weakest topics.
- Use only these marks. Don't guess why I lost marks.
- Give me a 5-day revision plan, 2 hours a day, with the most time on the weakest topics.

Why these lines

One question per line, with its topic, is what lets it add up a topic that appears more than once. "Marks out of that topic's total, with a percentage" matters because raw marks mislead: Linked lists scored 5, the same as Arrays, but out of 10, not 5. Ranking "by percentage" puts the weak topics at the top for the right reason. "Use only these marks. Don't guess why I lost marks" stops it inventing reasons it can't know from a list of numbers. And the plan line says how much time you really have: 5 days, 2 hours a day.

How it was tested

One run on 9 November 2026, on free ChatGPT in a web browser, not signed in, in a new chat, so there was no account, memory or saved chat history to draw on. The model in use wasn't shown. The six checks, the sample marks, the right answers and the prompt were written down and saved before the prompt was sent.

1. Every topic total is right: Arrays 5/5, Linked lists 5/10, Stacks 4/5, Queues 5/5, Trees 3/10, Hashing 5/10, Sorting 4/5, and 31/50 overall if it gives one.
2. Every percentage is right: 100, 50, 80, 100, 30, 50 and 80.
3. It ranks by percentage, not raw marks: Trees weakest, then Linked lists and Hashing (tied), and no topic scored 5 out of 5 is put level with them because it also has 5 marks.
4. The 3 weakest topics it names are Trees, Linked lists and Hashing.
5. Only the marks: it doesn't invent reasons for lost marks or say what the questions were about.
6. The plan is 5 days of 2 hours, 10 hours in all; Trees gets the most time, and Linked lists and Hashing each get more than any topic at 80% or above.

What came back

The sums were right, every one: each topic's marks out of its own total, every percentage, and 31 out of 50, 62%, overall. It ranked by percentage: Trees at 30% first, then Linked Lists and Hashing at 50%, with a note: "Topics with equal percentages share the same performance level." Arrays, also on 5 marks but out of 5, came out strongest, not level with the weak two. The 3 weakest it named were Trees, Linked Lists and Hashing, which is right. It said "No assumptions about why marks were lost", and kept to that.

The days of the plan were sound too. Five days of 2 hours, 10 hours in all. Trees gets all of day 1 and an hour of day 2, about 3 hours. Linked Lists and Hashing get about 2 hours each. Stacks and Sorting get half an hour each on day 4, and day 5 is mixed revision and a short timed mock. The study tasks it named for each day, like practising tree traversal, are general revision advice, not claims about what the mock asked.

Five of the six checks passed.

The catch

The miss is one box. Right under "Total study time: 10 hours (2 hours per day)." the answer has a "Time allocation" box: Trees 4 hours, Linked Lists 3, Hashing 3, Stacks & Sorting 2, Arrays & Queues 2. The box prints no total, but its five lines add up to 14 hours, not 10. And it doesn't match its own days: the days give Trees about 3 hours, the box says 4.

It's an easy one to miss, because the box is the neatest part of the answer, with coloured bars, and it sits right under the line that says 10 hours. Plan your week from the box and you've planned four hours that don't exist. Follow the days instead, and add up any summary before you trust it.

The stamp

Half works · 9 Nov 2026 · free ChatGPT, in a browser · 5 of 6 checks passed. It's good at turning a marked mock into a weak-topic list: right totals, right percentages, a ranking by percentage, the right 3 weak topics, and no guessing. The 10-hour day plan is sound. But its time box adds up to 14 hours under a 10-hour heading. Add up the hours yourself. One run is one run: try it with your own marks and see what yours does.

Before you use it

For revising only. Never in an exam, and for marked work, follow your school's or college's AI rules. ChatGPT is for ages 13 and over, or the minimum age in your country, and under 18 needs a parent's or guardian's permission ^[1].

Source 1: Terms of Use, Registration and access: minimum age (OpenAI, effective January 1, 2026) · openai.com · read 5 Nov 2026

The raw answer, word for word

Taken word for word from the page text. Its tables are written one row per line, with " | " between the cells; its coloured labels are in square brackets; its bar charts are written as "name · value". Nothing else changed.

1. Topic-wise marks and percentages

You attempted 10 questions, each worth 5 marks, for a total of 50 marks.

Topic | Marks obtained | Total marks | Percentage
Arrays | 5 | 5 | 100%
Linked Lists | 5 | 10 | 50%
Stacks | 4 | 5 | 80%
Queues | 5 | 5 | 100%
Trees | 3 | 10 | 30%
Hashing | 5 | 10 | 50%
Sorting | 4 | 5 | 80%
Overall | 31 | 50 | 62%

2. Topics ranked from weakest to strongest

Rank | Topic | Percentage
1 | [Weakest] Trees | 30%
2 | Linked Lists | 50%
3 | Hashing | 50%
4 | Stacks | 80%
5 | Sorting | 80%
6 | Arrays | 100%
7 | Queues | 100%

Topics with equal percentages share the same performance level.

3. Your 3 weakest topics

1. Trees · 30% · 3/10 marks
2. Linked Lists · 50% · 5/10 marks
3. Hashing · 50% · 5/10 marks

These are your three lowest-performing topics based strictly on the marks provided. No assumptions about why marks were lost.

4. Your 5-day revision plan

Total study time: 10 hours (2 hours per day).

The largest share of time goes to Trees, followed by Linked Lists and Hashing.

Time allocation
Trees · 4 hours
Linked Lists · 3 hours
Hashing · 3 hours
Stacks & Sorting · 2 hours
Arrays & Queues · 2 hours

Day 1 – Trees (2 hours)
[Highest priority]
45 min – Revise tree concepts and terminology.
35 min – Practise binary tree and BST problems.
30 min – Practise tree traversal questions.
10 min – Attempt a short self-test.

Day 2 – Trees and Linked Lists (2 hours)
60 min – Trees: practise problems and revise concepts.
45 min – Linked Lists: revise singly, doubly and circular linked lists.
15 min – Solve linked list practice questions.

Day 3 – Linked Lists and Hashing (2 hours)
60 min – Linked Lists: practise operations and problem-solving.
50 min – Hashing: revise hash functions and collision handling.
10 min – Quick self-test.

Day 4 – Hashing, Stacks and Sorting (2 hours)
60 min – Hashing: practise problems.
30 min – Stacks: revise and solve questions.
30 min – Sorting: revise algorithms and practise questions.

Day 5 – Full revision and mock test (2 hours)
30 min – Arrays and Queues revision.
30 min – Stacks and Sorting revision.
30 min – Mixed practice covering Trees, Linked Lists and Hashing.
30 min – Take a timed mini mock test and check your answers.

5. Final performance summary

Metric | Result
Overall score | 31/50
Overall percentage | 62%
Strongest topics | Arrays, Queues
Topics at 80% | Stacks, Sorting
Topics needing improvement | Trees, Linked Lists, Hashing

My recommendation: Prioritise Trees first. It is your lowest-scoring topic at 30%, and it should receive the most revision time. Then focus on Linked Lists and Hashing equally.

Your goal for the next mock should be to improve your performance in these three topics while maintaining your existing strengths in Arrays and Queues.

Next in the series, part 5: "The night before: a one-page summary".