Where to spend the last 180 minutes: a mock score profile

He is right about the marks and wrong about the inference, which is the most common error in exam preparation: highest value is not the same as highest gain.

Part 4 is the only part where items carry two marks, so six items yield 12 marks - more than any other part in this paper. But the two marks are all-or-nothing for the complete transformation. His 4/12 is the proof: two fully correct items, three near-misses, zero credit for the near-misses. Partial knowledge converts to nothing here.

Part 4 is also the slowest part to improve. Each item requires control of a specific structure - passive, conditional, reported speech, comparison, verb plus preposition - plus the discipline to hold a fixed frame while changing two things at once. Three weeks buys perhaps two more items.

Compare Part 3. Eight marks, one mark each, and rule-governed: identify the word class the gap needs, then apply the suffix or prefix pattern. Moving from 3/8 to 6/8 in three weeks is a realistic target, and it is the same skill that feeds Part 1 and Part 2.

The answer to give Dani directly: yes to a targeted slice of Part 4 - the high-frequency frames - and no to 'all of it'. His instinct to chase marks is correct; his calculation of where marks are recoverable is not.

Allocation rule: marks available × headroom ÷ time needed to learn.

• 60 min - Part 3 word formation (3/8). Three 20-minute cycles: nouns from verbs (-ance, -ment, -ion), adjectives (-ive, -ous, -able), negatives (un-, in-, im-, dis-). Each cycle: rule, eight items, error log. Highest headroom, fastest to teach. Target: +3.
• 45 min - Part 4 (4/12). Only the six recurring frames: too…to / so…that, passive, reported speech, conditionals, not as…as, verb + preposition. Teach the two-change rule: the given word usually forces both a structural and a lexical change. Target: +2 items.
• 30 min - Part 2 open cloze (4/8). A closed list: articles, prepositions in fixed phrases, linkers, pronouns, auxiliaries. Rule: the answer is almost always a grammar word, rarely a content word. Target: +2.
• 25 min - Part 6 (3/6). Reference chains and paragraph direction, as a class demonstration rather than individual practice. Target: +1.
• 20 min - Part 7 (6/10). Timed scanning; stems paraphrase the text, so match meaning not words.

Zero minutes on Parts 1 and 5: already strong, headroom is small, and time there is time taken from Parts 3 and 2. Dani should be shown this arithmetic on paper - it is the same reasoning he was already trying to do, done with better inputs.

Sitting one every lesson is not possible - 75 minutes plus review consumes a whole session - and the diagnostic value is now low. He already has a profile from this mock. Repeated full papers would mostly re-measure what is already known and crowd out the teaching that would change it.

Do two.

1. Start of week two, to check whether the Part 3 and Part 2 work transfers under time pressure.
2. Final week, full conditions including the answer sheet and a clock, to rehearse timing rather than to test knowledge.

Between them, use part-level timed practice - Part 3 in eight minutes, Part 7 in twelve. That builds the same exam skill at a fraction of the cost and gives immediate feedback while the reasoning is still fresh.

Two practical points to set now:

• Time discipline. Reading Parts 5–7 carry 22 marks between them; Part 4 carries 12. He must not let transformations eat the reading time. Decide the split in advance and leave the paper when the clock says so.
• Do not sit a full paper in the last lesson. A short, confidence-building set plus a review of his error log is a better use of the final 45 minutes than a dispiriting late score that changes nothing.

Finally, tell him why: exam technique is built by fixing specific parts, and a full paper is a measuring instrument, not a training exercise.