Skip to content
AbilityBench

Timed aptitude format

Abstract reasoning test on one clock across the whole set

Twenty figural items — eight matrices, seven series and five odd-one-out — with twelve minutes for all of them rather than a limit on each, free and with nothing held back at the end. That is the shape of a graduate aptitude test and the reason it feels different from an untimed puzzle: the budget is yours to spend, an item you linger on is charged to the items at the end, and the set stops where the clock stops.

  • 100% free
  • No signup
  • 20 items, 12 minutes
  • One budget, not per item
  • Three formats mixed

20 items in 12 m 00 s8 matrices, 7 figure series and 5 odd-one-out items, mixed. That budget is 36 seconds an item on average, and the average is the only thing that is fixed: you may spend two minutes on one item, and it comes out of the ones at the end.

Before the clock starts

  • There is no going back to an item, which is how most computer-delivered aptitude tests are built. Deciding is part of the task.
  • Skipping is offered and it scores zero. This set is not negatively marked, so a guess is worth more than a skip on every item — the report counts what your skips cost.
  • The clock is charged in frames actually delivered to this tab. If you switch away, the item on screen is voided by the engine and the seconds you were gone are not billed to you.
  • Answer with a click or with the number keys. Matrices have eight choices, series six, odd-one-out five.

The three untimed items are one of each type, with no clock running and nothing counted. The countdown does not begin until you press the button on the screen after them.

How to sit a set with one budget

The tactics are half the test, and they are the half nobody explains beforehand.

  1. Fix a per-item budget before the first item, then hold it

    Twelve minutes across twenty items is thirty-six seconds each. That is an average, not a rule, so the working method is to allow yourself roughly a minute on anything that looks solvable and to bail out of anything that does not. The failure mode this format punishes is not being slow, it is being slow on one item: four minutes spent on a matrix you eventually solve costs you the three items at the end that you would have solved in one minute each.

  2. Recognize which of the three formats you are looking at

    A 3x3 grid asks which tile completes it and gives eight candidates. A row of five figures asks what comes sixth and gives six. Five figures side by side with no arrow ask which one does not belong and give five. Each has a different chance level, and the odd-one-out items are the fastest of the three because they need one shared property found rather than a rule reconstructed.

  3. Guess rather than skip, then read the pace column

    This set carries no penalty for a wrong answer, which is normal for the format, so a skipped item is a guaranteed zero and a guessed one is worth between a fifth and an eighth of a mark. The report prints how many you skipped and what those skips gave away. It also prints your median time on each of the three formats, which is where a pacing problem shows up as one format eating the budget the other two needed.

Technical specifications

Set and budget20 scored items in 12 minutes, one countdown across the whole set. That is 36 seconds an item on average, spent however you like, and the countdown is visible from the first scored item onward
Item mix8 matrices (one and two and three rules per item), 7 figure series across four rule families, 5 odd-one-out items — shuffled with at most three of the same format in a row so the run never settles into one shape
Answer choices per format8 for a matrix, 6 for a series, 5 for an odd one out. Guessing every item blind therefore averages 3.2 marks out of 20, and the results panel prints that figure beside your own
NavigationForward only. There is no review screen and no flagging, which is how most computer-delivered aptitude tests are built — the decision to move on is part of what is being measured
SkippingOffered on every item and scored as zero. The report counts your skips and states what they would have been worth guessed, because no page should offer a control without saying what it costs
The clock and a lost tabThe countdown advances on frames actually delivered to this tab, so time spent with the page in the background is not charged. The item that was open when you left is voided by the engine rather than marked wrong, and a replacement joins the end of the set
Untimed run first3 items, one of each format, with no countdown running and nothing counted. The clock does not start until you press the button after them
What leaves the pageNothing. Marks, pace and the per-format table are computed from records held in this tab, and the copy button writes counts rather than answers to your own clipboard

Frequently asked questions

Is abstract reasoning the same as inductive or diagrammatic reasoning?

In publishers' catalogues the three names overlap heavily, and which one appears in your invitation email tells you less than you would hope. Abstract reasoning and inductive reasoning are usually the same figural task under two labels; diagrammatic reasoning is more often the one with process flows and operators in it, and is more common for software and engineering roles. The safe assumption when you are told to expect one of them is that you will see shapes changing according to a rule you have to infer, and that the differences between the three names are marketing rather than construct.

What counts as a good score on twenty items in twelve minutes?

Nothing does, and the reason is structural rather than modest. A cut-off would have to come from somewhere, and the only somewhere available is this page's own choice of how hard to make twenty items — move one rule and the same person scores three marks higher. What a publisher sells is not the item bank but the comparison group behind it, and the same raw score converts to different percentiles against graduate applicants, against managers and against one employer's own applicant pool. The figure worth carrying away from a practice run is not the count but the pace: how many items you finished before the clock, and whether the ones you lost were lost to the rule or to the time.

Should I really guess instead of leaving an item blank?

On a set with no negative marking, always. A blank is worth exactly zero and a blind guess is worth the reciprocal of the number of choices, so across a run the arithmetic is unambiguous. The only situation that reverses it is formula scoring, where a fraction of a mark is deducted for a wrong answer to cancel the value of guessing — that is rare in graduate testing and, where it exists, the instructions say so explicitly. If the instructions do not mention a penalty, there is not one.

Real assessments give more than twelve minutes. Why is this one short?

Because twelve minutes is the shortest budget that reproduces what the clock does to you, and a short run is one you will actually take twice. Pacing pressure appears as soon as the average per item drops under about a minute; it does not get more instructive at twenty-five minutes, it just costs more of your evening. What a longer sitting does add is fatigue, which is a real component of a long assessment and is deliberately not measured here.

Why can I not go back to an item I rushed?

Because the version of this format you are likely to meet does not let you either, and building in a review screen would train a habit the real thing punishes. Forward-only delivery is the norm for computer-administered aptitude sets, partly to stop candidates working the whole paper backwards from the easy items and partly because the item is meant to be answered from what is on the screen. If you want to go back over your reasoning, the untimed sibling pages keep every item for review at the end.

Does the countdown keep running if my machine stutters?

It runs on frames the browser actually delivers to this tab, so a stutter of a few dropped frames costs you those milliseconds and nothing more, and a switch to another window costs you nothing at all — the item you were on is thrown away instead. That is more generous than a real administration, which typically keeps counting and often reports the switch to the employer. It is the right trade for a practice page: the number here should describe your reasoning, not your notifications.

Are these the same items as the untimed pages in this set?

They come from the same generator with the same five construction rules, and every run draws fresh items from a new seed, so you will not meet an item twice unless you reuse a seed deliberately. That matters more than it sounds: an item you have already solved measures recall rather than reasoning, and a practice site that served everyone the same twenty items would be measuring how many times you had visited it.

What the clock is actually measuring, and what a publisher does with the number

The distinction that matters here is between a power test and a speeded one. A power test gives you as long as you want and asks whether you can solve the item at all; a speeded test gives you items you could all solve and asks how many you get through. Almost every commercial aptitude test is a hybrid, and the mixture is a design choice rather than an accident: a pure power test cannot be administered at scale, and a pure speed test would rank a room of graduates on clerical throughput. A twelve-minute budget over twenty items sits where most of them sit, which means the score you get here moves with two different things at once, and separating them is exactly what the pace column in the report is for. If your accuracy is high on the items you reached and you reached twelve of twenty, the clock beat you; if you reached all twenty and marked eight, it did not.

The second thing worth knowing before an assessment is what happens to your raw score after you hand it in. Nobody reads “fourteen of twenty”. The publisher converts it into a percentile against a comparison group chosen for the role, and the sift threshold — if there is one — is set on that percentile. This is why candidates comparing scores in a forum reach no conclusion: the same fourteen is a different result against a general population group, a graduate group and an employer’s own applicant pool for that specific job, and the group is picked by the employer rather than by the test. It is also the reason this page prints a count, a pace and a chance level, and stops. There is no population norm for “a verbal reasoning test” or “a pattern recognition test”, because those name a question format rather than an instrument. The score depends entirely on how hard the page made its own items, so any average quoted would describe this page's difficulty setting and nothing about people.

Which leaves the part of preparation that does transfer: knowing the formats cold so the budget is spent on rules rather than on working out what the screen is asking. The three here are the three that turn up. If you want the same figural material without the clock, so you can see whether the rules are there at all, that is the matrices page and, for the error breakdown, the matrix reasoning diagnostic. The series items on their own, with the rule-discovery cost timed, are the inductive reasoning test. A full graduate battery normally pairs the figural section with a verbal reasoning section and, increasingly, with a personality questionnaire that is not a test at all and is not scored like one.

A reaction time here is the interval between the frame that painted the stimulus and the timestamp the browser attached to your key, both read from the same monotonic clock. What neither can see is the display pipeline behind it, so on a 60 Hz screen roughly 16 ms of every figure below is the machine rather than you. That is the timing floor: two numbers closer together than that are the same number, and this page reports no precision it cannot support.

The countdown itself inherits that floor twice over — once at the frame that paints an item and once at the frame the clock is read on — so treat the last second of the twelve minutes as approximate. Everything above the second is exact, and the item durations behind the pace column come from the same paint-to-click measurement as the rest of this site.

This is a measurement exercise, not a clinical assessment. It reports what you did on this page against a stated reference and nothing more — it cannot establish the aptitude a particular employer is screening for, or a fit for the job behind it. Only a qualified professional, working with more than a browser, can make that judgment.

Where the twenty items and the twelve minutes are kept

Every number on this page is worked out by JavaScript running in the tab you are reading it in. Your answers, your reaction times and your score are never uploaded, logged or kept — which is also why the test carries on working after you disconnect from the network, and why nothing here can be held back behind an email address.

The countdown is a number in this tab and stops existing when you close it, which is also why an interrupted set cannot be resumed tomorrow. Nothing about your pace, your skips or your marks is recorded anywhere, and no employer has any way to see that you took this.