The engine behind the edge
Pinnacle is not a question bank with a shuffle button. Here is exactly how it finds your gaps, schedules your reviews, and tells you — precisely — when you are ready.
How we find your gaps
Most candidates don't know what they don't know. Pinnacle resolves that in the first 20 questions.
Every answer you submit sends five signals: topic, sub-topic, difficulty, time on question, and outcome. Those signals are written into your performance profile in real time.
After each session, Pinnacle computes a mastery estimate for every topic and sub-topic you have touched. The estimate is not a simple percentage correct. It weights harder questions more heavily, penalises apparent guesses implied by very low time on question, and decays confidence for topics you have not visited recently.
The result is a live Learning Profile that updates continuously — not just once at mock exam time. You see exactly which topics are at risk. The system often knows before you do.
The FSRS scheduling engine
Memory decays. The question is whether it decays before or after your exam.
Hermann Ebbinghaus showed in 1885 that without reinforcement, you forget roughly half of what you learn within 24 hours and 70% within a week. The solution is not to study more — it is to study at the right moments.
Spaced repetition exploits a feature of memory called the spacing effect: reviewing a concept just before you are about to forget it produces dramatically stronger retention than reviewing it again immediately. Each successful recall earns a longer interval before the next review.
Pinnacle uses FSRS v5 — the same algorithm Anki adopted after large-scale empirical testing showed it outperforms the older SM-2 algorithm on real retention data. SM-2, used by most platforms, systematically over-spaces easy material and under-spaces hard material. FSRS models your personal forgetting curve using four parameters: stability, difficulty, retrievability, and lapses.
Intervals grow faster for easier material and slower for harder material. Each candidate's schedule is unique to their own history. Fast-correct answers (strong recall) earn longer intervals than slow-correct answers (fragile knowledge).
How mastery is scored
A mastery score is not a percentage of questions answered correctly. It is a confidence estimate.
Specifically, mastery is a probability: given this candidate's history on this topic, how likely are they to answer a representative exam-level question correctly?
That probability factors in four variables:
- Question difficulty — a correct answer on a hard question moves the estimate more than a correct answer on an easy one. Easy questions only demonstrate easy recall.
- Recency — mastery estimates decay over time for topics not recently practised, reflecting real forgetting. A topic you aced six months ago is not a safe topic today.
- Consistency — alternating right and wrong answers on the same concept signals low genuine mastery even at 50% accuracy. The system detects and penalises this pattern.
- Volume — a single data point earns low confidence. Estimates tighten as you answer more questions. You earn certainty, not just a number.
An 80%+ mastery score on hard questions means exam-ready on that topic. Below 60% is the danger zone — Pinnacle surfaces those topics first in every session until the number moves.
The score is a tool for focus, not just measurement. Every session plan is built around moving the lowest scores up.
The 85% Challenge Point
Questions that are too easy produce no learning. Questions that are too hard produce frustration. Cognitive load theory identifies an 85% success rate as the optimal challenge zone — roughly 15% failure per session — where learning is maximised. Pinnacle tracks per-question difficulty empirically and routes each candidate toward questions calibrated to sit right at their personal 85% threshold. Your challenge zone shifts as you improve. The system shifts with you.
Ready to let the engine find your gaps?
Find my weak spots →
No credit card · First session ready in 60 seconds