Where GCSE Maths Marks Actually Leak
Ask examiners where students lose marks and the answer is rarely "they didn't know the maths". The marks that decide grades leak through a small set of repeatable errors: a dropped sign, a number rounded two steps too early, working done in the head, a question answered that was never asked. None of them are knowledge gaps. All of them are habits, which is the good news, because habits can be trained out.
The ten mistakes below cover the overwhelming majority of avoidable mark loss at GCSE. For each one you will see a concrete example of how it happens and, just as importantly, how the mark scheme punishes it, because understanding exactly which marks disappear is what makes the fix stick. Mark schemes award marks in pieces: method marks for a correct approach, accuracy marks for correct values, and final-answer marks. Different mistakes destroy different pieces, and knowing which is which changes how you check your work.
As you read, be honest about which of these appear in your own marked work. The practise-mark-fix loop only works if the "fix" step names the real error, and for most students two or three of these ten account for nearly all of their dropped marks.
- Most dropped marks come from habits, not knowledge gaps
- Mark schemes award method, accuracy and final-answer marks separately
- Each mistake destroys a specific type of mark, so each has a specific fix
- Most students lose almost all their careless marks to just two or three patterns
1. Sign Errors in Algebra
The single most common slip in GCSE algebra is mishandling a negative, and the classic site is expanding a negative bracket. Faced with , a rushed student writes instead of the correct : the negative multiplied the first term and then got forgotten for the second. The same error family appears when subtracting a whole expression, where becomes instead of , and when substituting a negative value, where into gets typed as and comes back rather than .
The mark scheme's punishment depends on where the slip lands. In a multi-step question, one sign error usually costs the accuracy mark for that step and the final-answer mark, though follow-through can preserve later method marks. The nastier case is a "show that" question: since the target answer is printed on the paper, an examiner seeing a sign error knows any correct conclusion was reverse-engineered, and the marks collapse.
The fix is mechanical, not motivational. Write every step, one operation per line, so the negative has nowhere to hide. Put brackets around every negative number before substituting, without exception: cannot be misread, can. And when you expand a negative bracket, say "minus times minus is plus" as you write the second term. It feels childish and it works.
- $-3(x - 4) = -3x + 12$: the negative multiplies every term
- Always bracket negative numbers before substituting: $(-2)^2$, never $-2^2$
- One sign error typically costs the accuracy and final-answer marks
- In "show that" questions a sign error can collapse the whole question
2. Rounding Too Early
Premature rounding is the quiet killer of multi-step questions, especially in trigonometry and geometry. Suppose a two-stage problem needs the hypotenuse of one triangle to feed into a second calculation. The true intermediate value is 7.6157... but the student writes 7.6, carries that forward, and the final answer lands outside the tolerance the mark scheme allows. The method was perfect. The answer is still wrong.
Mark schemes are explicit about this: final accuracy marks come with a required precision, often three significant figures, and an answer outside the accepted range does not score the accuracy mark, however sound the working. When an early rounding causes the miss, you keep your method marks but lose the accuracy marks, and on a question worth four or five, that is often two marks gone for no mathematical error at all. Questions that say "give your answer to 3 significant figures" are also flagging that rounding anywhere earlier is dangerous.
The fix has two parts. First, carry full calculator values between steps: use your calculator's answer memory rather than retyping a shortened number. If you must write an intermediate value down, keep at least four or five figures and mark it as unrounded. Second, round exactly once, at the very end, to exactly the precision the question asks for. If it does not specify, three significant figures is the safe convention, but write the fuller value above it so the examiner can see your unrounded answer too.
- Round once, at the end, to the precision the question asks for
- Carry full calculator values between steps using the answer memory
- An answer outside the scheme's tolerance loses the accuracy marks
- "Give your answer to 3 s.f." is a warning that early rounding will hurt
3. Not Showing Your Working
GCSE mark schemes award method marks independently of the final answer, and this cuts both ways. A student who shows a correct method but slips on arithmetic still collects most of the marks. A student who does the work in their head, or entirely inside a calculator, and writes only a final answer has bet the whole question on that one number. If it is right, fine. If it is wrong, the question scores zero, because there is no visible method for the examiner to credit.
Concretely: a 4-mark simultaneous equations question might award one mark for a correct elimination or substitution step, one for solving to find the first variable, one for finding the second, and one for the pair of answers. A student who writes only ", " and is wrong scores 0 out of 4. A student who shows all the steps but makes one arithmetic slip near the end typically scores 2 or 3 out of 4. Over a whole paper, that difference is a grade. Some question types raise the stakes further: anything worded "show that" or "prove" awards marks only for the argument, so an unexplained answer scores nothing even when it is correct.
The fix is to treat working as the product, not the scrap. One step per line, equals signs aligned, and every calculator computation written down before you key it in. Our dedicated guide to method marks in GCSE Maths breaks down how schemes allocate them question type by question type.
- Method marks are awarded independently of the final answer
- A wrong answer with full working often scores most of the marks
- A wrong answer with no working always scores zero
- "Show that" and "prove" questions award marks only for visible reasoning
4. Misreading the Question, and Answering the Wrong One
Two closely related mistakes live here, and between them they cost more marks than any algebra error. The first is misreading: taking in the numbers but not the conditions. A question says a ladder is 5 m long and reaches 4 m up a wall, and asks how far the base is from the wall; a rushing student computes the hypotenuse of a triangle that already had one. Or the question says "per month" and the student works in years, or gives values in centimetres and asks for the answer in metres.
The second is answering a different question from the one asked. The algebra is flawless, the student solves correctly, and then stops at and when the question asked for the sum of the solutions, or the perimeter the solutions describe, or which value is valid in context. Mark schemes are unforgiving here: the final-answer mark is for the answer to the printed question, and "answered a plausible neighbouring question" earns whatever method marks were passed through along the way and nothing more. On a 1- or 2-mark question, that usually means zero.
The fix costs about ten seconds per question. Underline the command words and conditions as you first read: the units, the required form, words like "hence", "exact", "to 1 decimal place", "in its simplest form". Then, after finishing, reread only the final sentence of the question and check your answer is literally the thing it requests. That final-sentence check catches a remarkable share of these errors, because the last sentence is almost always where the actual demand lives.
- Underline units, conditions and command words on first reading
- Watch for "hence", "exact", "simplest form" and stated units
- Solving correctly but answering the wrong demand forfeits the answer marks
- After finishing, reread the question's final sentence and check the match
5. Wrong Units and the Wrong Calculator Mode
Units first. Many mark schemes reserve a mark for the final answer being in the right units, or attach the accuracy mark to a correct value with correct units. Write 48 when the answer is 48 cm² and the scheme may still give it; convert wrongly, or mix units mid-question, and the damage is worse. The classic trap is a compound problem that gives a speed in km/h and a time in minutes: a student who multiplies them directly gets a number that means nothing, and every mark after that point is gone. Conversion errors between area and volume units are similar: 1 m² is 10,000 cm², not 100, and schemes routinely test exactly that.
Calculator mode is the same mistake wearing a different hat. GCSE trigonometry uses degrees, and a calculator left in radians answers every trig question wrongly: should be , but in radians mode the display shows , a value that should immediately look wrong for an angle in a triangle. The scheme cannot rescue this, because the method mark may survive but every computed value and the final answer are outside tolerance, so the accuracy marks all fail together, potentially across several questions before you notice.
The fixes are pure routine. Check the small "D" for degrees on your calculator display at the start of every paper, and again after any battery change or reset. Convert all quantities into consistent units before computing, writing the conversion as a line of working so it earns credit. And attach units to your final answer every single time; it is a free habit that protects a recurring mark.
- Convert to consistent units before calculating, as a written step
- 1 m² = 10,000 cm²; area and volume conversions are routinely tested
- Check the degrees indicator at the start of every calculator paper
- In radians mode, every trig value and accuracy mark fails at once
6. Not Checking Whether the Answer Is Plausible
Every answer in GCSE Maths lives in a context that tells you roughly how big it should be, and students who ignore that context hand marks away silently. A mean age of 340 for a class of students, a probability of 1.4, a ladder leaning at 90.3°, the price of a jumper after a 20% discount coming out higher than it started: each of these is an alarm bell that a whole checking step would catch in five seconds. Probabilities live between 0 and 1. Sides of triangles are shorter than the hypotenuse. Discounted prices go down.
The mark scheme does not award a specific "plausibility mark", which is exactly why this habit is undervalued: its value is that it triggers the re-check that recovers marks you had already lost. An implausible answer almost always traces back to one of the earlier mistakes in this list, a sign error, a units mix-up, a misplaced decimal from early rounding, and the student who notices the absurdity gets the chance to find and fix the slip. The student who writes down a probability of 1.4 and moves on has converted a recoverable error into a permanent one, and examiners' reports mention exactly these answers year after year.
Build the check into your rhythm: before leaving any worded question, ask "is this the right kind of size, and the right kind of thing?" Estimate crudely where you can. If a question multiplies 19.6 by 5.1, the answer should be near 100, and a calculator answer of 999.6 means a keying slip. Five seconds per question, several marks per paper.
- Probabilities must lie between 0 and 1; check every one you write
- Ask whether the size and type of the answer fit the context
- An implausible answer is a free prompt to find a recoverable slip
- Crude estimation catches calculator keying errors instantly
7. Time Management and Leaving Blanks
AQA and Edexcel maths papers carry 80 marks in 90 minutes, and OCR's carry 100 marks in longer sittings, so across the boards you have very roughly a minute per mark with a little slack for checking. The mistake is not slowness; it is spending that time in the wrong places. A student who sinks 15 minutes into a stubborn 3-mark question has paid five times its value, and the currency comes out of the accessible questions at the end of the paper that they now never reach. Papers ramp in difficulty, but not perfectly: there are almost always gettable marks late on, and the mark scheme pays the same for a mark earned in question 2 as in question 22.
Leaving blanks is the same error in its terminal form. A blank question scores zero with certainty, and it is the only score an examiner cannot argue with. Yet on worded and multi-step questions, the first mark is often cheap: writing the relevant formula, extracting the right numbers from the text, drawing the diagram, or computing an obvious first step. Mark schemes award these opening method marks even when the attempt goes nowhere, so an attempted question's realistic floor is not zero.
The fix is a rule decided before the exam, not during it. Give each question roughly its marks in minutes; if you are stuck beyond that, circle it, leave it, and move on without guilt. Sweep the whole paper first for marks you can bank, then return to circled questions with the time that remains. And in the final minutes, make sure nothing is blank: a formula, a first step, a sensible attempt on everything.
- Budget roughly a minute per mark and enforce it per question
- Circle and move on: return only after banking the accessible marks
- First method marks are cheap; an attempt's floor is rarely zero
- Never leave a question blank; a blank is the only guaranteed zero
Turning This List Into Marks
Ten mistakes make an intimidating list, but the practical version is much smaller: your marked work will show that two or three of them account for nearly everything you drop. The way to find out is to mark your own past papers strictly against the scheme and label every lost mark with its cause, sign error, early rounding, missing working, misread, wrong demand, units, mode, plausibility, time, or blank. After three papers the pattern is unmistakable.
Then train the specific fix, not general "carefulness". Carefulness is not a plan; a bracket around every negative number is a plan. So is a units-conversion line at the top of every compound-measures question, a final-sentence reread before moving on, and a mode check when you first pick up your calculator. Each fix is small, mechanical and quick, which is precisely why it survives exam pressure when good intentions do not.
The habits only become automatic through repetition under realistic conditions. Practising with instant marking helps because the error surfaces while the attempt is still fresh in your mind: Exam Ladder generates practice questions with worked solutions across all 70 GCSE topics and grades full mock exams against real exam-board grade boundaries, so a recurring careless error shows up in your results as a pattern rather than a mystery. However you practise, close the loop the same way: find the habit, name it, and drill its mechanical fix until the exam version of you does it without thinking.
- Label every dropped mark on three past papers to find your pattern
- Train mechanical fixes, not general carefulness
- Two or three of the ten mistakes will explain most of your losses
- Repetition under timed conditions is what makes the fix automatic
Frequently asked questions
What is the most common mistake in GCSE Maths exams?
Not showing working is the most expensive habit, because it turns partial knowledge into zero marks. GCSE mark schemes award method marks independently of the final answer, so a wrong answer with clear working often scores most of a question while a wrong answer alone scores nothing. Among pure mathematical slips, sign errors when expanding negative brackets and substituting negative numbers are the most frequent.
Do you lose marks for not writing units in GCSE Maths?
Often, yes. Many mark schemes attach the final mark to a correct value with correct units, or reserve a mark for units on measurement questions. Mixing units mid-question is worse still, because a calculation combining metres with centimetres or hours with minutes produces a wrong value and loses the accuracy marks that follow. Convert everything to consistent units as a written first step and attach units to every final answer.
How much working do I actually need to show?
Enough that an examiner can follow your method without guessing: one step per line, including the calculation you are about to type into your calculator. You do not need to narrate in sentences. The test is whether someone could award the method marks from the page if your final answer were covered up. On "show that" and "prove" questions the working is the answer, so every logical step must appear.
What should I do if I am stuck on a question in the exam?
Give a question roughly its marks in minutes; beyond that, circle it and move on. Marks later in the paper cost the same as marks earlier, and the accessible ones must be banked first. When you return, write something rather than nothing: the relevant formula, the numbers extracted from the text, or a first step. Opening method marks are awarded even for incomplete attempts, so a blank is the only guaranteed zero.
How do I stop making careless mistakes in maths?
Replace the vague goal with specific mechanical habits. Mark your recent papers, label every dropped mark with its cause, and you will find two or three recurring patterns. Then drill the matching fix: brackets around negative numbers, rounding only at the final step, a calculator mode check at the start, a reread of the question's final sentence before moving on. Practised under timed conditions, these become automatic within a few weeks.