IELTS Band 8 preparation

    From Band 7 to Band 8: the error-elimination guide

    Band 8 is not more English — it is fewer recurring errors. Census your patterns, eliminate them one by one, and confirm with mocks.

    • Band 8 numbers per paper
    • Error patterns that cap 7
    • 4-week protocol

    In short

    Moving from Band 7 to Band 8 means converting recurring errors into occasional ones: Band 8 describes a very good user whose mistakes are rare and unsystematic. The work is therefore a census, not a course — log every error across mocks, rank patterns by frequency, and eliminate them one per week. This guide gives the per-paper numbers, the patterns that most often cap candidates at 7.x, and the four-week protocol; SpeakMetry's mocks with per-criterion corrections supply the census data.

    The last mile: what changes between 7 and 8

    Same skills, different error rate. That reframes the whole task.

    1. Band 7 rewards strong performance with visible flaws: developed ideas, flexible vocabulary, varied grammar — plus a recurring article slip, a memorised linker in every paragraph, a distractor that keeps working. Band 8, a very good user with full operational command, keeps the strengths and loses the patterns: errors become occasional and unsystematic rather than habitual. Nothing new must be learned; several small things must stop happening.

    2. This is why broad study stalls at 7.5. Another vocabulary list cannot fix an article system, and another full course cannot fix a timing habit — only a census of your own errors, ranked by frequency, tells you which two or three patterns own your missing half band. Candidates who log every loss for two weeks almost always discover the gap is narrower, and more specific, than it felt.

    3. Consistency is the second half of Band 8. One 35 in Listening proves ability; three consecutive 35+ sittings prove the band. Productive papers round down, so a Writing profile of 8, 8, 7.5, 7.5 averages 7.75 and lands at 7.5 — near-clean criteria, repeatedly, is the actual requirement. Plan for confirmation mocks, not just peak performances.

    The numbers for Band 8

    Exact thresholds from the official conversion tables — no folklore.

    Listening: 35 correct

    Thirty-five out of forty, meaning at most five losses across four parts — spelling, limits and distractors included. Misses must be rare, not merely fewer.

    Academic Reading: 35 correct

    Same bar on harder texts: 35 with the clock obeyed. GT Reading asks 37, since everyday texts are expected to yield more.

    Productive criteria near-clean

    Writing and Speaking average whole-band criteria and round down — 8, 8, 7.5, 7.5 becomes 7.5. One lagging criterion vetoes the section band.

    Overall math is unforgiving

    An 8.0 overall needs a 7.75 average: 8.5 + 8 + 8 + 7.5 works, while 8 + 8 + 7.5 + 7.5 lands at 7.5. There is no room for a weak paper anywhere.

    Writing decides most 8s

    Listening and Reading yield to drills fastest; Writing's criterion mix moves slowest without feedback. Budget repair time accordingly — the essay is usually the bridge or the wall.

    Error patterns that cap 7

    Fix patterns, not sentences: each pattern below, eliminated, upgrades dozens of future answers. Illustrative pairs, not verdicts.

    Grammar patterns

    System slips that repeat in every essay and talk.

    1. Articles with generics: “Government should invest…” → “Governments should invest…” or “The government should invest…”. One rule, applied everywhere.
    2. Agreement in complex sentences: errors hide in long subjects — “The list of reasons for studying abroad are…” → “…is…”. Check the head noun, not the nearest one.
    3. Comma splices: two full sentences joined by a comma. Split them or subordinate one; examiners count every splice.

    Lexical patterns

    Approximation that Band 7 tolerates and Band 8 does not.

    1. Collocation slips: “do a mistake”, “big rain” → “make a mistake”, “heavy rain”. Keep a personal collocation log from your own corrections.
    2. Memorised linkers: “moreover” opening three paragraphs → varied natural links, or none where the logic is already clear.
    3. Repeated vague nouns: “things”, “aspects” doing structural work → precise terms from the topic itself.

    Exam patterns

    Technique losses that survive every amount of English study.

    1. Listening transfer slips: right answer, wrong spelling or format on the answer sheet. A two-minute transfer routine ends them.
    2. Task 2 second idea underdeveloped: first body paragraph full, second thin. Plan both ideas with examples before writing either.
    3. Word-limit casualties in Reading and Listening completion: the answer inside extra words still scores zero. Count every time.

    These are illustrative practice prompts written in the official format, not official IELTS questions. They show you the shape of each part so you can rehearse out loud before your mock.

    The 4-week elimination protocol

    One pattern per week, confirmed by mocks. Nothing else changes.

    1. 1

      Week 1: census

      Sit full mocks of each paper and log every single error with its pattern name. Rank patterns by frequency — the top two own your missing half band.

    2. 2

      Week 2: pattern one

      Drill only the top pattern: targeted exercises, rewrites of old answers, re-recorded talks. All other study is maintenance. Confirm with a single-paper mock.

    3. 3

      Week 3: pattern two

      Same treatment for the second pattern while guarding the first — old patterns resurrect the moment attention leaves. Old mocks re-marked show whether they stayed dead.

    4. 4

      Week 4: confirmation

      Two full mocks at target across all papers, reviewed for pattern recurrence rather than marks. Then taper: the protocol ends, the habits continue.

    8.0 readiness signals

    Book when most of these hold across consecutive sittings — not after one good day.

    • Receptive mocks consistently at 35+ correct, not occasionally
    • No productive criterion below 7, most at 8, across sittings
    • The same error does not appear in three consecutive mocks
    • Full mocks clear the target twice in a row with margin
    • Timing finishes comfortable with minutes to spare, not exactly on zero
    • The error log shrinks week over week instead of refilling

    Explore more ways to prepare

    Frequently asked questions

    What does IELTS Band 8 mean?
    It describes a very good user: full operational command of English with only occasional unsystematic inaccuracies. In numbers, about 35 correct out of 40 in Listening and Academic Reading, with productive criteria mostly at 8.
    How many correct answers is Band 8 in each paper?
    Band 8 starts at 35 correct out of 40 in Listening and Academic Reading, and at 37 in General Training Reading. Writing and Speaking have no raw marks — they need near-clean criterion scores, since section bands round down.
    Why am I stuck at Band 7 or 7.5?
    Almost always recurring patterns rather than missing ability: an article system, a collocation habit, a timing leak, an underdeveloped second idea. A two-week error census names them precisely — and named patterns are fixable where vague “improve English” is not.
    How long does it take to move from 7 to 8?
    It depends on how many patterns own the gap and how consistently you drill them — often weeks of focused pattern work rather than months of courses. Your own mocks answer this honestly: shrinking logs mean approaching readiness, static logs mean the diagnosis is wrong.
    Which paper should I fix first?
    The one furthest from its Band 8 number: raw marks short in receptive papers, the lowest criterion in productive ones. Writing is the usual decider, so it typically deserves the largest share of pattern weeks.
    Can mock tests measure Band 8 readiness?
    Receptive mocks measure it exactly through raw marks. Productive mocks estimate it through per-criterion bands plus corrections on your own sentences — and repeated estimates trending upward across weeks are the most reliable readiness signal available outside the real test.
    Where does the data for the error census come from?
    SpeakMetry's simulators provide the census data: full mocks of every paper with per-criterion feedback and corrections tied to your answers. Your first sitting of each paper is free, and a one-time pass covers the daily drills and confirmation mocks.

    Next step

    Census your errors this week

    Two mocks, one ranked log, and the two patterns that own your missing half band.

    Find your error patterns