Nine Forms, Three Families, No Shared Questions

A single form has an obvious problem: the second time somebody takes it, they are partly being tested on their memory of it. That is not a marginal effect on a reasoning set, where recognising that you have seen a matrix before is most of the work.

So the challenge ships parallel forms. Each family follows one blueprint, and the forms within it contain entirely different questions. Retaking a parallel form is an honest second attempt in a way that retaking the same one is not.

• Quick Challenge: 18 scored questions in about 12 minutes, the fastest way to try the format.
• Full Challenge: 54 scored questions in about 36 minutes, with broader mixed coverage of all four domains.
• Language-Light Challenge: 36 scored questions in about 24 minutes, emphasising matrix, sequence, spatial and symbol-rule items with short prompts.

Language-Light reduces reading load. Prompts are short, and the item types lean toward patterns and symbols rather than verbal reasoning. For somebody working in a second language, or anybody who finds long prompts a tax rather than a test, it is a materially different experience.

It is not culture-free, and it is not language-free. No test is. Claiming otherwise is one of the most common overstatements in this field, and the pages here say so in as many words rather than letting a name imply it.

The reason the claim is impossible is that the conventions a reasoning item relies on are learned somewhere. Reading a matrix left to right and top to bottom, treating a rotation as a transformation rather than a different shape, and expecting a sequence to continue rather than repeat are all cultural habits of instruction, not universal facts about minds. A shorter prompt removes reading load. It does not remove any of that.

The Practice Lab is the opposite of a form: single item types, drilled one at a time, untimed, with the reasoning explained as you go. Working through a few modules before a first form means the timer is testing your reasoning rather than your unfamiliarity with the layout.

That is not a trick for inflating a score. Format familiarity is a genuine confound in timed reasoning, and removing it makes the result more informative rather than less.

Starting a form creates a session with a fixed deadline. You answer one question at a time, can mark any question for review, and can return to marked questions before submitting. The in-progress attempt saves to this browser as you go, so an interruption on this device does not lose your work.

One honest limit: the deadline is enforced on the server, so the clock keeps running through an interruption. Treat a form as one sitting, and use the untimed routes when that is not possible.

Marking for review is worth using properly rather than as a way of postponing everything. The productive pattern is to answer with your best guess, mark it, and move on, so that running out of time costs you a revision rather than an answer. Leaving a question blank to come back to it is the version that ends badly.

Parallel forms only work if the blueprint is specific enough to constrain them. Within a family, the forms hold the same number of scored items, the same distribution across the four reasoning domains, and the same timing, so the thing that changes between them is the questions and not the shape of the test.

That is what makes a second attempt interpretable. If two forms differed in length or in domain mix, a difference in your result would tell you nothing: it could be the forms rather than you. Holding the blueprint fixed is unglamorous work and it is the only thing that makes comparison across attempts meaningful at all, which is also why the aptitude suites record the version that scored an attempt.

Start with a Quick form unless you already know you enjoy this. Twelve minutes is enough to find out whether the format suits you, and the Full forms are a real commitment of attention rather than time alone.

If reading load is the obstacle rather than the reasoning, start with Language-Light instead, and do not treat the two as interchangeable measures of the same thing.