The 60 percent rule
On most speeded cognitive tests, answering 60 percent of the questions with 90 percent accuracy beats answering 100 percent of the questions with 70 percent accuracy. The math is straightforward. On a 50-question test, the first scenario gets you 27 correct. The second gets you 35 correct, but only if you believe you can actually sustain 70 percent accuracy under panic pace, which most candidates cannot.
Internalize this early. You are not trying to answer everything. You are trying to maximize the expected count of correct responses. Those are not the same objective.
The 10-second triage
Within the first 10 seconds of reading a question, you should have a verdict. Either this is a question I can solve comfortably in under a minute, or it is not. The 10-second window exists because your gut is a remarkably accurate classifier once you have drilled question families.
If the answer is yes, commit and solve. If the answer is no, flag and move. Staying to prove something to yourself is ego, not strategy. Ego is expensive on a timed test.
Pacing in thirds
Mentally divide the test into three equal thirds by time, not by question count. The first third is for momentum. Pick off the questions you recognize immediately and bank confident points. The middle third is for careful work on medium-difficulty questions. The final third is where you convert leftover time into guesses on flagged questions and blind commits on the final minute.
The thirds rule prevents the classic pattern of working too carefully in the first half and running out of clock at the end. Checking your progress against the third markers is a cheap way to catch drift before it becomes a disaster.
The flag-and-return technique
Nearly every modern aptitude test lets you flag questions. Flagging is not a sign of weakness. It is how you turn the test interface into a prioritized queue for whatever time remains at the end.
Practice flagging aggressively. If a question is not resolved within its time budget, flag it and commit a guess immediately rather than leaving it blank. A committed guess is better than a blank because you may not have time to return.
What 509 timed sittings actually show
Most pacing advice, including most of the advice above, is reasoning from first principles. We can do better than that, because every question sat on PrepClubs records what was answered and how long it took. The numbers below come from 696 re-graded attempts, 509 of them on the four tests with a genuinely punishing clock: CCAT, Wonderlic, PI Cognitive and Cubiks Logiks.
Three things fall out of the data, and all three contradict how most candidates behave.
PrepClubs Research, 2026-09-21. 727 attempts analysed from 1,473 on record, re-graded against the live question bank, attempts using under half the clock excluded. Method at /research/method.
A blank is the only answer guaranteed to score zero
Nearly three quarters of candidates leave something empty, and the median candidate leaves 8 questions blank. None of these tests penalises a wrong answer, so every one of those blanks is a guaranteed zero where a guess had a real chance.
Filling them in at random would be worth about 2.1 raw points to the average candidate, which is roughly 7 percent on top of a mean score of 28.3. That is not a rounding error on a test where a few points move you a percentile band.
And guesses made under pressure are not actually random. Accuracy does fall as the clock runs down, from 84.8 percent in the first fifth of the test to 53.3 percent in the last, but 53.3 percent is still more than two and a half times the 20.7 percent you would expect from picking at random. Even your rushed answers carry real knowledge. Leaving the box empty throws that away.
| Where you are in the test | Answers given | Correct |
|---|---|---|
| First fifth | 5,081 | 84.8% |
| Second fifth | 5,047 | 72.1% |
| Third fifth | 4,607 | 70.7% |
| Fourth fifth | 3,448 | 60.4% |
| Final fifth | 2,119 | 53.3% |
| Pure chance | 20.7% |
The answer count falls across the test because most candidates never reach the end, which is the subject of our other study. Accuracy is measured only on questions that were actually answered.
The 50-second question is what costs you the end of the test
The median question on a speeded test takes 14 seconds. Nine in ten are done inside 39 seconds. Only 5.8 percent run to 50 seconds or beyond, so a long question is genuinely unusual rather than normal.
Those rare questions are expensive. Candidates who never spent 50 seconds on a single item answered 93.8 percent of their test. Candidates where at least one question in twenty ran to 50 seconds or more answered 65.3 percent. That is a gap of nearly thirty percent of the paper, and the questions you never reach are not the hard ones, they are simply the last ones.
Be careful how you read that comparison, because we are not going to overclaim it. It is observational, not an experiment: people who sink 50 seconds into a question may be slower or weaker across the board, not only worse paced. What the data does establish is that a long question and an unfinished test travel together, reliably, across 339 attempts.
The practical rule follows: at around 50 seconds, answer with your best guess and move. Not because the question is unwinnable, but because the two or three questions waiting at the end of the paper are just as likely to be easy, and you will never see them.
How to guess intelligently
Guessing is not random. Eliminate the option you are most certain is wrong. On multiple choice with five options, eliminating one bumps your odds from 20 percent to 25 percent. Over 50 questions, those expected-value gains add up to one or two extra correct answers.
Many tests use distractor options that are deliberately tempting. Learn to spot them. On numerical questions, watch for answers that result from a single misapplied operation. On verbal, watch for answers that use a word from the passage but in the wrong sense. Both are classic distractors.
Clock discipline without obsession
Check the clock once every 10 questions, not once every question. Constant clock checking fragments attention and slips you into anxiety. Once per checkpoint is plenty. Pair each check with a quick, honest assessment: am I on pace, ahead, or behind? Adjust immediately rather than hoping it averages out.
If you are behind at the midpoint, shift into rapid-elimination mode. Do not panic, but also do not try to make up the gap with careful work on hard questions. The best recovery is confident pace on easy and medium questions you can still clear.
Rehearsing pacing in practice
Pacing is a muscle. It will not show up automatically on test day if you have not trained it. In practice, set a strict timer and commit to the thirds rule from session one. Do not practice without the timer and tell yourself you will turn it on next session. You will not.
Record how many questions you attempt in each third. Over a week, you will see your per-third counts even out and your accuracy rise. That is the signal that pacing has become reflex rather than intention.