Ask a Year 11 class where they lose marks and you get the same answer every time: silly mistakes. Careless errors. Rushing. The assumption behind that answer is that the marks are lost in ones, scattered thinly across a paper that was mostly going fine.
We were able to check. Marky marks whole scripts against the official mark scheme and records every question part separately, so we pulled 155 GCSE maths scripts and counted where the marks went, using Marky’s own marks.
The answer is not what students think it is.
The headline
Across those 155 scripts there were 4,670 separately marked question parts, carrying 12,020 marks between them. Students earned 7,428 of those marks, which is 61.8%.
Split by how each part was marked:
- 56.1% of parts scored full marks. More than half the paper, done and correct.
- 25.4% scored nothing at all.
- 18.4% scored something but not everything.
That middle number is the one worth sitting with. A quarter of all the question parts attempted came back with a zero next to them.
And when you count the actual marks lost rather than the parts, it gets starker. 67.9% of every mark lost in those 155 scripts came from a question part that scored zero. Only 32.1% came from parts where the student got some credit and missed the rest.
So the picture students carry around is close to backwards. Marks are not mainly trickling away one at a time from questions that nearly worked. They are going in blocks, from questions that produced no credit whatsoever.
Where the blocks are
Not every question loses marks at the same rate, and the pattern is clear once you group parts by how many marks they were worth.
| Marks available | Question parts | Marks lost | Share of all marks lost | Share scoring zero |
|---|---|---|---|---|
| 1 mark | 972 | 278 | 6.1% | 28.6% |
| 2 marks | 1,326 | 731 | 15.9% | 21.0% |
| 3 marks | 1,401 | 1,711 | 37.3% | 26.1% |
| 4 marks | 662 | 950 | 20.7% | 21.8% |
| 5 marks | 309 | 922 | 20.1% | 39.5% |
| All parts | 4,670 | 4,592 | 100% | 25.4% |
Percentages are rounded to one decimal place, so the highlighted column adds up to 100.1 rather than exactly 100.
Read the highlighted column downwards. The 3, 4 and 5 mark questions together account for 78.1% of every mark lost, despite being only about half of all the parts on the paper.
The single biggest line is the 3 mark question. On its own it swallows 37.3% of all lost marks. It is the most common part on a GCSE maths paper and it is the one students are least likely to fully land.
The 5 mark question is the most brutal per attempt: 39.5% of them scored zero. Two in five of the longest questions on the paper produced nothing at all. Not a partial answer, not a method mark. Nothing.
What a zero on a 5 mark question actually means
Here is the thing that should change how you revise. A GCSE maths question worth 5 marks is almost never awarded as one lump. The mark scheme breaks it into steps, and the steps are awarded separately. A typical one might read:
- M1 for a correct method to find the area of the cross section
- A1 for 42
- M1 for multiplying by the length
- M1 for converting the units
- A1 for the final answer
Five marks, five separate opportunities, and they are not all-or-nothing. You can get the final answer completely wrong and still walk away with three of those five, provided the examiner can see the method you used.
Which means a zero on a question like that usually is not a wrong answer. It is a blank space, or working the examiner could not follow. The student who wrote nothing and the student who wrote a full page of correct method with an arithmetic slip at the end score very differently, and only one of them scores zero.
That is why the 5 mark questions have the worst zero rate on the paper. They are the ones students look at, decide they cannot do, and leave.
The near miss is real, but it is small
There is a second pattern in the data and it is worth naming honestly, because it is the one students already believe in.
683 question parts, 14.6% of the total, lost exactly one mark. Those are the genuine near misses: the answer left as an unevaluated expression, the unrounded decimal, the missing unit, the negative sign that vanished on the last line.
They are real and they are worth fixing. Fourteen per cent of the paper is not nothing. But they account for a small slice of the total damage, and if you spend your revision entirely on being neater and more careful you are polishing the part of your performance that was already nearly fine.
The bigger prize is the question you did not start.
What to do about it
Four changes, in rough order of how many marks they are worth.
1. Never leave a multi mark question blank. This is the single highest value habit in GCSE maths and it costs you nothing. If a question is worth 4 marks and you can do the first step, do the first step and write it down. The mark scheme is looking for that step whether or not you ever reach the answer. On our numbers, the questions students abandon are where two thirds of all lost marks live.
2. Write the method even when you are unsure it is the right method. Method marks are awarded for a correct process, not for a correct outcome. An examiner cannot award a method mark for a method that only ever existed in your head or on your calculator screen. Which lines actually earn that credit is worth knowing precisely: what “show your working” means to an examiner.
3. Practise the 3 mark question specifically. It is the most common part on the paper and the largest single source of dropped marks. Three marks usually means two steps and an answer, or a method, an intermediate value and a conclusion. Learn the shape of it.
4. When you check a past paper, count your zeros before you count your total. Your total tells you a grade. Your zeros tell you what to do on Monday. Ten scattered dropped marks and ten marks lost in two abandoned questions produce the same score and require completely different responses. This changes how a past paper is worth doing at all: most students use them wrong.
The honest limits of these numbers
These are 155 real GCSE maths scripts marked against the official schemes, and they are a genuine sample of student work rather than anything we made up. They are not a national sample. They are the scripts that happened to be marked by Marky, from schools that shared class sets with us, and they are one subject at one level. The counts are Marky’s marks, not a teacher’s. A cohort with a different profile would produce different percentages.
What we would stand behind is the shape rather than the precise figures: lost marks in GCSE maths concentrate in longer questions, and they concentrate in the ones that scored nothing. We would expect that to hold well beyond this sample, because it follows from how the mark schemes are built.
For teachers reading this
The reason this analysis was possible at all is that the marking was recorded question by question rather than as a total. A pile of marked scripts with a percentage on the front of each contains all of this information and surrenders none of it.
When Marky marks a class set, every part is stored separately with the marks won, the marks available and the reasoning, so this same table exists for your cohort by the evening of the test. Which is when it is still worth acting on, and which is most of the case for not marking everything by hand.