AR: Speed Strategies & Building Your Pattern Library
The 14-Second Challenge
AR is the most time-pressured subtest. You have about 14 seconds per question. For Type 1 questions, you need to identify both set rules within ~30 seconds, then classify each test shape in ~10 seconds. This is only achievable through pattern recognition — not through systematic analysis of each box.
Building Your Pattern Library
The fastest AR performers have a mental library of common patterns. They don’t analyse from scratch each time — they recognise familiar pattern types instantly. Build your library by:
- After every practice session, write down every pattern/rule you encountered
- Categorise them: counting rules, shape rules, colour rules, position rules, conditional rules, relationship rules
- Review your library before practice sessions to prime your brain
- The more patterns you’ve seen, the faster you’ll recognise new ones
Speed Heuristics
- Start with colour/shading: It’s the most visually obvious property — check it first
- Then check count: Quickly count the total number of shapes in each box
- Then check shape types: Is a particular shape always present or absent?
- Last resort — position and conditional rules: These are hardest to spot quickly
When to Guess and Move On
If you cannot identify the rule within 20 seconds, your best strategy is:
- Make your best guess for each test shape (use partial pattern matching if you’ve spotted something)
- Flag the question set
- Move on to the next set
- Return only if you finish early (unlikely in AR, but possible)
Avoiding Common AR Mistakes
- Don’t look for overly complex rules — most UCAT AR rules are simple. If your proposed rule involves 3+ conditions, you’re probably overcomplicating it
- Don’t focus on one box — always verify against ALL 6 boxes in the set
- Don’t ignore the distractor elements — knowing what’s irrelevant is as important as knowing what’s relevant
- Don’t spend too long on one set — 5 quickly-answered questions from the next set are worth more than 1 perfectly-answered question from a hard set