What adaptive scoring means in practice
The test is estimating your ability and refining that estimate. Early answers move the estimate most, because there is least information. So the pacing instinct people bring from paper tests — go fast early, slow down for the hard ones later — is backwards here.
The exact mechanics differ by test and by section, and both test makers publish detail on it. Read theirs rather than a forum's — this is a domain where confident folklore vastly outnumbers accurate information.
What actually raises a score
- Quant is a fundamentals test with a costume on. The maths is not advanced; the questions are engineered so that the obvious approach is slow. Learning the fast approach is most of the score.
- Verbal is largely vocabulary and reading, both of which respond to [spaced repetition](/tools/spaced-repetition-app) — and vocabulary in particular is a pure memory problem that people insist on solving with word lists they read once.
- Timed, full-length, sat in one go. Endurance is a real component and it cannot be trained in twenty-minute chunks.
- Error logs, categorised — was it a gap, a timing failure, a misread, or a slip? Only one of those is fixed by more content review.