Mental Math for Game Assessments: How to Get Fast Under a Timer

Every HireVue-style battery includes at least one numerical game, and rapid arithmetic is the purest of them: a stream of questions like 54 − 48 or 7 × 13, a short timer, and a score built from how many you clear correctly. Nothing here is beyond primary-school maths — the test is entirely about speed under pressure, which is trainable in a way "intelligence" is not.

What the game actually measures

Employers read this as numerical fluency: how cheaply your brain retrieves number facts and runs small calculations. Fluency frees attention for the actual decision — the same reason a fluent reader understands a paragraph better than someone sounding out words. The scoring almost always combines throughput (questions answered) with accuracy, and like the other games, an error usually costs more than a slow correct answer.

The five techniques that buy the most speed

• Left-to-right, not column arithmetic. For 47 + 38, do 40 + 30 = 70, then 7 + 8 = 15, answer 85. School's right-to-left carrying method is optimised for paper; left-to-right gives you the big digits first and finishes faster in your head.
• Round, then repair. 54 − 48 is 54 − 50 = 4, +2 = 6. 9 × 99 is 9 × 100 − 9 = 891. Working near round numbers halves the effort of most questions the game throws.
• Check the units digit first. Endings must match: 13 × 7 must end in 1, so if your candidate answer ends in anything else you have caught the error before typing it. In pair-finding games like Numerosity, endings-first scanning is the core strategy — a target ending in 7 needs digit pairs like 3+4 or 8+9.
• Re-memorise the times tables you have lost. Most adults have gone rusty on 6×, 7×, 8× facts and pay a half-second toll every time one appears. Ten minutes of drilling brings retrieval back to instant — the cheapest points available anywhere in the battery.
• Type the moment you know. Under a timer, the gap between "knowing" and "committing" is pure loss. Trust the ending-digit check and answer; second-guessing a correct answer into a timeout is one of the most common failure patterns.

Pacing: the boring secret

Fast candidates do not sprint. They hold a steady rhythm slightly below maximum speed — the pace at which they are almost never wrong — and let the clock do the rest. Bursts feel productive but produce error clusters, and adaptive difficulty means one sloppy patch drags the next questions down with it. The same metronome principle wins in Disco Numbers, where each round is a single magnitude comparison and rhythm is everything.

A one-week training plan

1. Days 1–2: drill multiplication facts 6–9 and complements to 100 (37 + ? = 100) until instant.
2. Days 3–4: play the mental math game focusing purely on accuracy — find the pace where you stop making errors.
3. Days 5–6: push that pace up gradually; add Numerosity for endings-first pair-finding.
4. Day 7: simulate test conditions — one sitting, no pauses, all three numerical games back to back.

Percentile scores after each run show you exactly whether the pace push is paying for itself.

Keep it in context

Numerical games are only one slice of the battery — the complete guide covers how memory, attention and spatial games sit alongside them, and the 10 tips guide has the cross-game habits (sleep, warm-up, environment) that quietly decide close percentiles. Start with the free mental math practice — five minutes tells you your baseline.

Numbers under a timer show up in more than one costume. Disco Numbers is the same arithmetic dressed as a memory task, and the pacing habits below transfer to it almost unchanged.

Where you'll meet this game

Timed-arithmetic tasks show up across many employers' assessments. If you're preparing for a specific process, J.P. Morgan's pymetrics games and HSBC's online assessment both include numerical challenges of this kind. For the provider's full battery and how numerical fluency is scored within it, see the complete pymetrics games guide.

What do people actually score?

First attempts here average out at a median of 16 correct, with the top tenth at 19, across 59 players. The median gain from first run to best is 25 percentiles, which is one of the strongest practice effects we can measure.

FAQ

How fast is fast enough? Fast enough that arithmetic is not the bottleneck — the question should take longer to read than to answer. Our own first-attempt data puts the middle of the range at 16 correct answers.

Am I allowed a calculator? Assume not unless your invitation says otherwise. Even where one is allowed, reaching for it on simple steps costs more time than doing them in your head.

Does estimating count as cheating? No — it is the skill being measured. Knowing roughly what the answer should be before you calculate is what separates people who finish from people who run out of clock.