How Your IELTS Writing Band Score Is Calculated and Rounded

Candidates often ask what percentage they need for band 7. The question has no answer, because IELTS Writing does not use percentages. There is no total to divide by, no pass mark, and no raw score out of 20 or out of 100.

What exists instead is a small piece of arithmetic that runs twice and then once more. It is completely public, and once you can do it yourself, two useful things happen. You stop guessing what your practice essays are worth, and you can see exactly which criterion is holding your score down.

This post walks through that arithmetic step by step, with real number combinations, and shows the single quarter band that decides more results than anything else.

Why there is no percentage

Reading and Listening are marked out of 40 questions, so a raw score exists there and gets converted to a band. Writing is different. Nobody counts your correct answers, because there are no answers to count.

Instead, an examiner reads your response and awards a whole band from 0 to 9 on each of four criteria. Those four whole numbers are the entire raw data. There is no denominator, so a fraction like 17 out of 20 does not translate into an IELTS Writing band at all. If a teacher has given you a mark out of 20, that is their own scale, not the IELTS one.

The four criteria

Task 2 is marked on these four:

1. Task Response. Did you answer the whole question, take a clear position, and develop your ideas?
2. Coherence and Cohesion. Is the essay organised, paragraphed, and easy to follow from sentence to sentence?
3. Lexical Resource. Range, precision and naturalness of vocabulary, including spelling and word formation.
4. Grammatical Range and Accuracy. Range of sentence structures, and how often errors appear and whether they interfere with meaning.

Task 1 uses the same last three. Its first criterion is called Task Achievement instead, because the job is different: report what the visual shows, cover the key features, and stay accurate.

Each criterion gets a whole band. You cannot receive 6.5 for Lexical Resource. Half bands appear only after the averaging step below.

Step one: average the four criteria

For each task, the four criterion scores are added and divided by four. Because you are averaging four whole numbers, the result can only ever end in .0, .25, .5 or .75. Nothing else is possible.

That average is then reported to the nearest half band. The .5 and .0 cases are already half bands and stay as they are. The two interesting cases are .25 and .75, and the published IELTS convention rounds both of them upward: .25 goes up to the next half band, and .75 goes up to the next whole band.

Here is every pattern you are likely to meet.

Read the second row carefully. Three sixes and one seven produce a reported 6.5. One criterion moved by a single band, and the reported score moved by half a band. That is the whole reason targeted work beats general practice.

Step two: combine the two tasks, with Task 2 counted twice

The Writing section has two tasks and one final score. They are not weighted equally. Task 2 is worth twice as much as Task 1, so the calculation is:

overall = (Task 1 band + Task 2 band + Task 2 band) divided by 3

That result is again reported to the nearest half band.

The doubling has a consequence people underestimate. Consider two candidates with the same two numbers in the opposite order.

Identical work, different distribution, half a band apart. If you have sixty minutes and you are behind schedule, the forty minutes belong to Task 2. This is also why finishing Task 2 matters more than polishing Task 1, and why an underlength Task 2 is so expensive.

A full worked example

Say your Task 1 report scores Task Achievement 6, Coherence and Cohesion 6, Lexical Resource 5, Grammatical Range and Accuracy 6.

Sum is 23. Average is 5.75. That ends in .75, so it rounds up to 6.0. Task 1 band: 6.0.

Your Task 2 essay scores Task Response 7, Coherence and Cohesion 6, Lexical Resource 6, Grammatical Range and Accuracy 6.

Sum is 25. Average is 6.25. That ends in .25, so it rounds up to the half band. Task 2 band: 6.5.

Overall: (6.0 + 6.5 + 6.5) divided by 3 is 6.33, which reports as 6.5.

Now change one thing. Suppose Lexical Resource in Task 2 had been 7 instead of 6. Task 2 sum becomes 26, average 6.5, band 6.5. The overall is unchanged. But if Grammatical Range and Accuracy in Task 2 had risen to 7 as well, the sum is 27, average 6.75, band 7.0, and the overall becomes (6.0 + 7.0 + 7.0) divided by 3 = 6.67, reporting as 6.5 still. It takes a Task 1 improvement to 6.5 to push the overall to 7.0.

That is the honest picture. Single criterion gains do not always move the reported number, which is why chasing the overall band is frustrating and chasing the weakest criterion is not.

What this means for your practice

1. Score your practice work by criterion, never as one number. Four numbers tell you where to work. One number tells you nothing.
2. Find your lowest criterion and work only on that for two weeks. For most candidates at band 6 it is Lexical Resource or Grammatical Range and Accuracy, because those two reward slow accumulation rather than exam tricks.
3. Protect Task 2 time. It carries double weight, and an unfinished Task 2 damages Task Response before the other three criteria are even read.
4. Stop converting to percentages. A band is a description of your writing, not a mark out of anything.

If you want to see this arithmetic run on your own essay, Bandly gives you the four criterion estimates and the rounded band, so you can see which single number is holding the result down. The scores are estimates produced by a language model, not marks from a qualified examiner, so treat them as a guide to where to work rather than a prediction of your result.