Capp-Style Assessment

What the real one is reported to be

Rothschild & Co says its online assessment is not timed and takes 30 to 45 minutes; the instructions candidates quote say 69 questions in one sitting. Candidates who sat it in the 2025 and 2026 cycles describe three blocks: about 14 situational judgement scenarios, about 40 behavioural sliders, and 12 numerical questions on charts and tables, one per page, easier than they expected. Cappfinity records the time on every answer and it feeds the result alongside accuracy.

Rothschild does not name its provider. The untimed, slider-based format is Cappfinity's, and some prep sites still describe an older, timed SHL version, so treat the provider as reported. One assessment is reused across applications in the same year, and after it comes a video or HR interview and then an assessment centre.

How it is scored here

A run is 44 items: 12 scenarios drawn from a bank of 20, all 20 slider pairs, and 12 numerical questions generated and checked before they are served. Each scenario and each numerical question is worth 10 points, so the most a run can score is 240. A ranking scores by its distance from the key: one adjacent swap costs two points and the reverse order scores nothing. A most and least pick scores five for each end, three when your pick is one place in. The sliders are never scored.

The scale is ours; no provider publishes one. It pays for a near miss because the ranking between two plausible answers is the skill. The run is capped at forty minutes and ends when the last item is answered.

Where we had to choose

The firm, Marlowe Hart, and the day are fictional and every scenario is written by us; nothing copies a published question. The key is simplytk's judgement, grounded in what Rothschild says it looks for (communication, drive, sound judgement, executional excellence, creativity, teamwork, numeracy) and the values it names (thoughtful, creative, principled), and it is explained item by item at the end of a run. The behavioural sliders have no key anywhere: the readout says where you sat and whether your two answers on a trait agreed, which is the thing Cappfinity is described as checking, and nothing more.

The numerical questions reuse the generator behind Numerical Reasoning in Cappfinity's three answer styles: pick, type and rank. Candidates describe the real ones as charts and tables and not hard; nobody confirms deal-style data, so the tables are ordinary business figures.

How to work through one

On the scenarios, rule out the two you would not do first, then decide which of the rest keeps the right person informed and leaves the decision with them; that one usually wins. Say the uncomfortable thing early and specifically. On the sliders, answer as yourself and consistently: the same trait comes round twice in different words, and a profile that contradicts itself reads worse than any lean. On the numbers, read the column header for the unit before you calculate, and do not rethink: the candidates who sat it say there are no trick questions.