SHL Numerical Reasoning Test: the complete 2026 guide

The SHL Numerical Reasoning Test is the most widely used numerical assessment in graduate and professional hiring. Forge has analysed 100+ SHL-style numerical questions and candidate reports to map which formats you will see, what the recurring traps are, and how to prepare efficiently. If you are new to all this, start with our psychometric tests overview; if you are not sure this is even the test you are facing, check which SHL test you are taking first.

Quick summary: SHL Numerical in 2026

• Three formats are live: Verify Interactive (10 questions / 18 min), Verify Numerical Ability (16 questions / 20 min), and Verify G+ (24 questions / 36 min).
• Maths level is GCSE: percentages, ratios, averages, tables and charts — no algebra or calculus.
• Forge’s database identifies five dominant question archetypes and four recurring trap types that account for most errors.
• Scores are reported as a percentile; large graduate schemes often set cut-scores around the 50th–70th percentile.
• There is no negative marking — never leave a question blank.

Which SHL Numerical format are you taking?

Check your invitation email before you prepare anything. Three formats are live in 2026:

A fourth format — 18 questions in 25 minutes — appeared until roughly 2020 and is now retired. If you see practice material referencing it, it is outdated.

The five question archetypes (from Forge’s database)

Forge’s SHL-style numerical database shows five recurring archetypes that together cover almost every numerical scenario you will face. Arithmetic budgeting alone makes up roughly 40% of questions and produces the highest error rate when gross and net figures appear in the same table.

1. Arithmetic budgeting (40% of SHL numerical questions)

You are given a budget, forecast, or variance table and asked how much was over/under spent, what percentage of budget was used, or what the revised figure is after an adjustment. The most common trap: the question specifies “net expenditure” but a distractor uses the gross figure, which is also visible in the same table.

2. Rate and ratio (20% of SHL numerical questions)

Two quantities, often in different units, combined to produce a rate — revenue per employee, output per machine-hour, cost per kilogram. The trap is per-unit normalisation failure: calculating the total correctly but forgetting to divide by the unit count, or dividing by the wrong denominator.

3. Percentage change (20% of SHL numerical questions)

Calculate the percentage increase or decrease between two figures, or apply a stated percentage change to derive a new figure. The most common trap: using the new value as the base rather than the original. If a figure rises from 200 to 240, the increase is 20% of 200 — not 16.7% of 240. Distractor options are calibrated to match the wrong-base calculation.

4. Table optimisation (10% of SHL numerical questions)

Find the row, combination, or constraint that maximises or minimises some quantity. Which supplier gives the lowest total cost per unit when delivery fee and minimum order are included? The trap: a visible “best” option in one column that becomes wrong when a second constraint is applied.

5. Chart-adjusted comparison (10% of SHL numerical questions)

A chart is paired with a table. The question requires reading a value from the chart, adjusting it using a figure from the table, then comparing across categories. The most common error: reading the chart without applying the adjustment.

Where candidates lose marks: trap type breakdown

Forge’s database tags every question with its primary trap type. Gross-vs-net and wrong-base percentage traps together account for about 60% of all incorrect answers — fixing just those two patterns typically adds 2–3 extra correct answers in a 10-question Interactive test.

The maths level

The arithmetic is GCSE-level. No algebra, no calculus, no statistics beyond percentages and averages. The test is hard not because the maths is hard, but because:

• Data is presented in dense tables with irrelevant columns designed to slow you down.
• Distractor options are calibrated to match the most common mistakes.
• The time limit creates pressure that causes calculation errors.

Calculator policy

SHL explicitly permits an on-screen calculator on all Verify formats. Use it for multi-digit multiplication and division. Do not rely on it for simple percentage calculations — if you need a calculator to find 15% of 800, you will consistently run out of time.

Mental arithmetic to build before test day:

• Percentages of round numbers: 10%, 5%, 25%, 20%, 50%, 75%
• Multiplying and dividing by small integers (2–12)
• Common fractions as percentages: 1/3 ≈ 33.3%, 1/8 = 12.5%, 3/8 = 37.5%

Scoring and what the employer sees

Your raw score is converted to a percentile against a norm group — recent graduates for graduate schemes, working professionals for experienced hires. The employer receives your percentile, not your raw score. Based on candidate reports and Forge’s benchmarking:

• 50th percentile: roughly 6–7/10 on Interactive, 9–10/16 on Numerical Ability
• 70th percentile: roughly 8/10 on Interactive, 12–13/16 on Numerical Ability
• 90th percentile: near-perfect accuracy with speed

There is no penalty for wrong answers on current SHL Verify formats. Never leave blanks.

Four failure modes to avoid

1. Practising without timed pressure

Doing questions untimed gives you a false sense of competence. Forge users who move from untimed to strictly timed practice typically see a 10–20% score drop at first — then recover to a higher plateau. From day one, practise with a clock running.

2. Missing the question qualifier

The question will say “net”, “excluding tax”, “in Q3 only”, or “after the adjustment in footnote 3”. Slow down for five seconds at the start of each question to identify the qualifier before touching any numbers.

3. Trusting the first matching distractor

If you calculate an answer and it matches option A, that is not confirmation you are correct — it may be confirmation you made the expected mistake. When time allows, verify by a second method or at minimum recheck your denominator and unit.

4. Calculator dependency slowing you down

Candidates who reach for the calculator on every step run out of time. Build fluency with percentages and ratios so calculator use is reserved for genuinely complex arithmetic.

7-day preparation plan

Forge’s free skill drills cover every archetype below — sign up once and they are included. Full SHL-style timed tests with exact time limits and answer explanations are available with a Forge plan.

• Day 1 — Baseline. Confirm your format from the invitation email. If you have a Forge plan, take a timed SHL-style test under real conditions. If not, run one timed drill per archetype and note which question types you miss.
• Day 2 — Percentage and ratio. 45 minutes on Forge’s free Percentage & Ratio drill (easy → hard). Do at least 20 questions. After every wrong answer, name the trap before moving on.
• Day 3 — Budget and variance tables. Forge’s free Table Interpretation drill. Practise finding the net figure in multi-column tables and resisting the gross-figure distractor.
• Day 4 — Charts. Forge’s free Bar Chart, Line Graph, and Pie Chart drills. Read chart values precisely and apply the table adjustment before comparing.
• Day 5 — Rate and ratio with unit tracking. Write down the unit you want in the answer before calculating. Flag every question where you used the wrong denominator.
• Day 6 — Full timed mock. Complete SHL-style test on Forge under exam conditions. Review the trap-type label on every wrong answer to identify your two weakest archetypes.
• Day 7 — Targeted review. 30 minutes on your two weakest archetypes from Day 6. 15 minutes on the free Estimation drill for mental arithmetic speed. Rest.

Related SHL guides

• SHL Tests: the complete provider hub
• SHL Verbal Reasoning Test
• SHL Inductive Reasoning Test
• SHL Deductive Reasoning Test
• Psychometric Tests 101

Frequently asked questions