The G-Research Quiz, Explained

G-Research — the London-based quantitative research fund — runs one of the most talked-about early screens in quant hiring: a timed mathematics quiz that candidates describe as short, sharp, and unforgiving. (As with all firm pages: this reflects consistent candidate reports, not official firm material, and formats evolve — treat it as orientation, not gospel.)

The reported format

• Timed multiple-choice mathematics — most commonly reported as 10 multiple-choice questions with a roughly 90-minute window (some candidates describe shorter, faster variants), so the constraint is sustained accuracy rather than raw speed. See the full G-Research online assessment guide for the stage-by-stage breakdown.
• Content mix: probability and expectation at the core, with combinatorics, sequences/series, quick calculus, and occasional logic. Difficulty ramps — early questions are warm-ups, late ones are genuinely hard under the clock.
• No calculator-friendly numbers. Answers tend to be exact forms; arithmetic fluency matters.
• Used early and decisively — it is a filter before serious interviewer time, which is why the pass bar is reputed to be high.

What it is filtering for

Speed-accuracy under time pressure on olympiad-flavored (not olympiad-hard) material. The quiz punishes two habits: slow-but-careful solvers who cannot triage, and fast pattern-matchers who misread twists. The winning mode is triage — bank the fast questions, commit remaining time to the hard ones, and guess strategically if unanswered questions cost nothing.

A practice plan that maps to it

• Timed probability sets. Drill the probability bank in test mode with a hard per-question clock — the material overlap is direct.
• Expectation fluency. The expectation bank (75 free) covers the quiz's center of mass.
• OA-style pacing. The timed OA simulations train the triage skill itself — knowing when to abandon a question is half the score.
• Mental math maintenance. A daily 80-in-8 or Zetamac session keeps the arithmetic layer frictionless.

Past the quiz, candidates report technical interviews on probability, statistics/ML, and coding — the firm funnels cover comparable research-fund loops stage by stage.