Acumen Logic®How it works
Diagnose. Teach. Prove.
Three steps, in that order. You find out where you stand, you learn the method behind everything you missed, and then you sit the real format and watch the number move. This page walks through each one with the actual mechanics.
Seventeen minutes, taken cold.
The Benchmark is 20 timed questions in three sections: numerical, verbal and logical reasoning. It is free, and so are the results: your score, an estimated percentile, your weakest section named, and the correct answer for every question.
You get one attempt, ever. Retake anything often enough and the number tells you about your familiarity with that test, not your ability. One attempt is what makes the result an honest baseline.
- Numerical: 8 questions, 8 minutes
- Verbal: 7 questions, 6 minutes
- Logical: 5 questions, 3 minutes
- Scored on our server against a fixed key
What you see at the end
Score, estimated percentile, weakest section.
Side by side, with the answer review underneath.
20 questions · 17 minutes · one attempt
- Numerical: 6 / 8
- Verbal: 3 / 7
- Logical: 4 / 5
Weakest section: Verbal, 3 of 7.
Step 02 · Teach
A drill is five units, not a pile of questions.
Every one of the 25 drills follows the same five units, in this order. Click one.
What a wrong answer gets you
Not "incorrect". A name, a method, and a line to remember.
This is the Try It unit of the Percentages & Percentage Change drill. Answer it, wrongly if you like, and watch the right-hand side.
After a 20% discount, a jacket costs £80. What was the original price?
What caught you Reverse percentage solved forwards
You applied the discount forwards instead of undoing it. A 20% cut leaves 80% of the original, so the original is the sale price divided by 0.8.
Step-by-step method
- £80 is 80% of the original
- Original = £80 ÷ 0.80
- = £100
Remember
To undo a percentage change, divide by the multiplier. Never add the percentage back.
The panels light up when you answer. Get it right and only the method and reminder show.
Named traps
The mistakes have names. Learn them.
Every wrong answer in a drill is tagged with the specific trap that caused it, with a worked example and what to watch for. Here are six. Click one.
Step 03 · Prove
Sit the real format. Then look at the numbers.
Sixteen replica forms across SHL, Aon and Cappfinity, built to each publisher's published specification. Every sitting feeds the analytics.
Numerical, verbal ×2, inductive, deductive.
Five options, about a minute a question, no penalty for wrong answers, no going back.
- Deductive Reasoning · 18 q · 20 min
- Inductive Reasoning · 18 q · 24 min
- Numerical Reasoning · 18 q · 20 min
- Verbal Reasoning — Test 1 · 30 q · 19 min
- Verbal Reasoning — Test 2 · 30 q · 19 min
Speeded, negatively marked, tabbed.
The real test deducts for wrong answers (publisher-confirmed). The replica scores 1 or 0 and explains where a guess would have cost you.
- Deductive Reasoning · 20 · 6 min
- Inductive Reasoning · 32 · 17 min
- Numerical Reasoning · 37 · 12 min
- Situational Judgement · 16 · 20 min
- Verbal Reasoning · 49 · 12 min
Untimed, time recorded.
Cappfinity does not publish item counts, so our set follows the format and scoring model it does describe, and says so.
- Critical Reasoning · 18 items
- Numerical Reasoning — Test 1 · 14 items
- Numerical Reasoning — Test 2 · 14 items
- Numerical Reasoning — Test 3 · 14 items
- Verbal Reasoning — Test 1 · 14 items
- Verbal Reasoning — Test 2 · 14 items
- Critical Reasoning: five dimensions; inference on a 5-point scale, the other four binary
Analytics
Accuracy, pace and estimated percentile, per publisher, per sitting.
Score history across every attempt. Category gaps ranked with a pace verdict. The traps that still catch you, from your drill answers. A plain recommendation of what to do next.
Last 10 sittings · accuracy by category
- Numerical: 62%
- Verbal: 76%
- Logical: 60%
- Situational: no data
Numerical: 62%, fast. You are rushing, not running out of time. 75% is where the engine calls a category solid.
Honesty
Every percentile says "estimated".
Your percentile is an estimate from a statistical model calibrated to typical performance on the aptitude tests UK employers set — not a live count of other people's scores. A model is not a measurement, so we label it, everywhere. How the estimate is produced →
Know where you stand
before they decide.
The Benchmark is free. Seventeen minutes, an instant estimated percentile, and the correct answer for every question.
Start Free Benchmark