Skip to main content

Field Guide · For medical students

Weak subjects from your own test marks

Plan your study (medical students), part 4: weak subjects found from practice test marks and revision hours split, tested on free ChatGPT.

· 4 min read

sources read up to

zenlite is independent of OpenAI. Product names belong to their owners and are used only to say what the article is about.

After a practice test the marks go into a notebook, and the hours go to whichever subject feels worst. But the subject with the best average can be the one slipping. This is part 4 of the Field Guide series on planning your study as a medical student: one job, tested on a real screen, with the result said plainly.

The job, and what to type

Part 3 dated the revision rounds. This part decides where the hours go. The marks below are a sample, made up for this demo: four subject names and three practice test scores each, nothing from inside a subject. Type your own marks in their place: subject names and numbers only, no name or roll number, never a classmate's, a class list or a photo of a mark sheet.

Two things make this sample harder than it looks. Test 2 was marked out of 50, not 100, so its raw numbers don't compare with the others. And Microbiology has the best average of the four while falling in every test.

The prompt, word for word:

I'm an MBBS student. Here are my marks from three practice tests. Tests 1 and 3 are out of 100. Test 2 is out of 50.
- Pathology: 48, 21, 51
- Pharmacology: 52, 28, 61
- Microbiology: 72, 32, 55
- Forensic Medicine: 40, 24, 58

Find my weak subjects and split my revision time:
- First turn every mark into a percentage.
- A subject is weak if its average is below 55%, or if its percentage dropped in every test.
- I have 20 hours a week for revision. Give weak subjects more hours. Every subject gets at least 3 hours. Whole hours only. The total must be exactly 20.
- Show a table: subject, the three percentages, average (one decimal place), trend (up, down or mixed), weak (yes or no), hours a week.
- Subject level only. No topic lists.

The 55% line, the 20 hours and the 3-hour floor are the sample's numbers. Use yours.

How it was tested

One run on 13 November 2026, on free ChatGPT in a web browser, not signed in, in a temporary chat: no memory, no past chats. The page didn't say which AI model answered. OpenAI's model release notes say its 29 October GPT-6 update applies to people using ChatGPT without signing in [2]. ChatGPT is for ages 13 and up, and under-18s need a parent's or guardian's permission [1].

Source 2: Model Release Notes: GPT-6 in ChatGPT update, October 29, 2026 (OpenAI Help Center) · help.openai.com · read 13 Nov 2026

Source 1: Terms of Use, Registration and access: minimum age (OpenAI) · openai.com · read 13 Nov 2026

Six checks were written down, with an answer key, before the prompt was sent:

  1. Every mark turned into a percentage (test 2 doubled) and every average right to one decimal: Pathology 47.0, Pharmacology 56.3, Microbiology 63.7, Forensic Medicine 48.7.
  2. Microbiology marked weak because it dropped in every test (72, 64, 55), even with the best average.
  3. Pathology and Forensic Medicine marked weak (average below 55%), Pharmacology not weak.
  4. Trends right: Pathology mixed, Pharmacology up, Microbiology down, Forensic Medicine up.
  5. Hours: whole numbers, each at least 3, exactly 20, every weak subject more than Pharmacology.
  6. Subject level only, and no health, sleep or medicine advice.

What came back

About 20 seconds later it began: "Since Test 2 is out of 50, its marks are multiplied by 2 to convert them into percentages." Then a table with every percentage, average and trend right, and the weak column: Pathology yes, Pharmacology no, Microbiology yes, Forensic Medicine yes. On Microbiology it said: "Weak because marks have dropped in every test, despite a good average."

The hours: Pathology 5, Pharmacology 4, Microbiology 6, Forensic Medicine 5, total 20. It ended by checking its own rules: more time for each weak subject than Pharmacology, at least 3 hours each, exactly 20. All 6 checks passed.

Worth a look

The most hours, 6, went to Microbiology, the subject with the best average, because it is the one falling. That's allowed by the rules in the prompt, and it is a fair reading. But it's one split among several that fit: the rules never said which weak subject should get the most. If Pathology at 47% worries you more, move an hour. The tool sorted the numbers; the hours are your call.

The stamp

Works · 13 Nov 2026 · free ChatGPT, not signed in · 6 of 6 checks. It handled the test marked out of 50, got every average and trend right, caught the subject slipping behind a good average, and kept every hours rule. One run is one run, on a made-up sample.

Before you start

For revising only. Never in an exam, and for marked work, follow your college's AI rules. Type your own marks and say what each test is out of. Keep it at subject level.

The raw answer, word for word

Taken word for word from the page. ChatGPT shows its tables as grids; here each table row is one line with " | " between cells, and the summary in section 3 is one item per line. Nothing else is changed. A screenshot of the page is kept with this post's test log.

1. Marks converted into percentages

Since Test 2 is out of 50, its marks are multiplied by 2 to convert them into percentages.

Subject | Test 1 | Test 2 | Test 3
Pathology | 48% | 42% | 51%
Pharmacology | 52% | 56% | 61%
Microbiology | 72% | 64% | 55%
Forensic Medicine | 40% | 48% | 58%

2. Weak subject analysis and weekly revision plan

Trend is based on the sequence of percentages across the three tests.

Subject | Test 1 | Test 2 | Test 3 | Average | Trend | Weak? | Hours/week
Pathology | 48% | 42% | 51% | 47.0% | Mixed | Yes | 5
Pharmacology | 52% | 56% | 61% | 56.3% | Up | No | 4
Microbiology | 72% | 64% | 55% | 63.7% | Down | Yes | 6
Forensic Medicine | 40% | 48% | 58% | 48.7% | Up | Yes | 5
Total | | | | | | | 20

3. Final weekly revision allocation

Total revision time: 20 hours
Pathology: 5 hours (25%)
Pharmacology: 4 hours (20%)
Microbiology: 6 hours (30%)
Forensic Medicine: 5 hours (25%)

4. Final assessment

Pathology: Weak because average is below 55%.

Pharmacology: Not weak. Average is above 55%, and performance is improving.

Microbiology: Weak because marks have dropped in every test, despite a good average.

Forensic Medicine: Weak because average is below 55%, although performance is improving.

Your three weak subjects are Pathology, Microbiology and Forensic Medicine.

The plan gives each weak subject more revision time than Pharmacology, maintains at least 3 hours per subject, and uses exactly 20 hours per week.

The test log

  • Run: 13 Nov 2026, free ChatGPT, web, not signed in, temporary chat, about 20 seconds; model not shown on the page. Checks and answer key written and committed before the run.
  1. Percentages and averages: pass. Test 2 doubled; 47.0, 56.3, 63.7, 48.7.
  2. Falling subject marked weak: pass. Microbiology, "despite a good average".
  3. Low averages weak, Pharmacology not: pass.
  4. Trends: pass. Mixed, up, down, up.
  5. Hours: pass. 5, 4, 6, 5, total 20.
  6. Subject level, no health advice: pass.

Stamp: Works, 6 of 6 checks.

Next in the series, part 5: the plan slipped, re-plan in 5 minutes.

Sources

Each source was opened and read on the date shown. Numbers in square brackets in the text point here.

  1. Terms of Use, Registration and access: minimum age (OpenAI)openai.com · read 13 Nov 2026
  2. Model Release Notes: GPT-6 in ChatGPT update, October 29, 2026 (OpenAI Help Center)help.openai.com · read 13 Nov 2026

General study information, not medical advice.