TacticsUpdated June 2026

Test-day tactics

Method gets you to a correct answer; tactics decide how many correct answers you have time to give. These six are where prepared candidates separate from merely capable ones — every one of them is trainable in days.

Pacing: decide your budget before the test decides it for you

Every test has an implied per-question budget — SHL gives you roughly 75 seconds, Saville 70, Talent Q 90 falling to 75, the speed tests under 20. Know yours before you sit down, and check the clock every three or four questions against where you should be. Pacing failures are rarely dramatic: they're 40 extra seconds, three times, and suddenly the last four questions get 15 seconds each.

The single most valuable habit: when you read a question and feel no plan forming within about ten seconds, mark it as a skip candidate. The question you wrestle with for three minutes costs you two easier marks elsewhere — a trade you should never accept.

The skip reflex

Skipping feels like failure, so most candidates do it too late. Reframe it: a skip is a trade of one uncertain mark for one or two near-certain ones. On speed tests it isn't optional — the CCAT and Wonderlic are built so almost nobody finishes, and the winners are the ones who decline the hardest questions fastest.

Practical rule: one read, one plan attempt, then either you're calculating or you're gone. If your test allows review, note skipped numbers; if it doesn't (most per-question-timed tests), pick the best remaining option after elimination and move on without looking back.

Estimate before you calculate

Before any precise calculation, get a rough answer: round 4,872 to 5,000, 23% to a quarter. Two reasons. First, on multiple-choice tests, the rough answer often eliminates three of five options immediately — and sometimes leaves only one. Second, it catches your own slips: when the calculator says 412 and your estimate said about 40, you've caught a misplaced decimal that would otherwise have matched a trap option perfectly.

The test writers know the common errors — wrong base, inverted ratio, missed unit conversion — and plant those exact wrong answers among the options. A matching option is not confirmation you're right; it's often confirmation the trap worked.

Calculator discipline

On calculator tests, the machine isn't the bottleneck — your hands are. Keep intermediate results on paper or in memory deliberately, not by retyping; chain operations where you can (× 1.12 in one step beats × 12 ÷ 100 + original in three); and never let the calculator replace the estimate that sanity-checks it.

On no-calculator tests, the numbers are chosen to be mentally tractable — if your working is getting ugly, you've usually taken a wrong turn. Look again for the shortcut: 5% is half of 10%, ×25 is ×100÷4, fractions of round numbers before decimals.

Read the exhibit before the trap reads you

Most wrong answers in data interpretation come from reading the wrong number, not computing the wrong way: the wrong row, the adjacent column, the units footnote (£000s!), the secondary axis on a combination chart. Spend your first seconds on the exhibit's structure — what the rows are, what the columns measure, what the units say — before touching the question.

In True/False/Cannot-say formats, the discipline is different: the trap is concluding more than the data states. 'Cannot say' is the answer whenever the statement needs information the exhibit doesn't contain — however plausible it sounds.

The final days

Two days out, stop learning new methods — the marginal mark is in pacing and rest, not another pattern. Run one timed mock or simulator per day at your real test's pace, review only the named mistakes, and sleep properly: reaction-speed losses from a short night are worth more dropped marks than any topic you might cram.

On the day: real test conditions are non-negotiable for your last practice run — the timer on, the calculator policy matched, no pausing. The version of you that shows up under pressure is the one you trained under pressure.

Tactics only stick under a real clock

Reading about the skip reflex changes nothing; skipping under a 19-second countdown three times does. Train these against your real test’s timing.

Common questions

Should I guess or leave questions blank?

On most numerical tests there's no negative marking, so an educated guess always beats a blank. Eliminate one or two options first — on a 5-option question that turns a 20% shot into 33% or 50%.

How long should I spend per question?

Whatever your test's budget says — roughly 60–75 seconds on SHL-style tests, under 20 on CCAT, Wonderlic or cut-e. Decide your per-question budget before you start and treat it as a hard rule.

Is it better to finish or to be accurate?

On speed tests (CCAT, Wonderlic), almost nobody finishes — keep moving and skip freely. On per-question-timed tests there's nothing to finish early for, so use the full budget. On whole-test timers, bank the easy marks first.