Every answer shows its working
The correct answer is never asserted at you. You get the mechanism, step by step, with the source behind each step — and each wrong option explained, with the reason it is wrong rather than just a cross.
Every question comes with the mechanism traced step by step, the sourced sentence behind each step, and — where it comes from a journal, a guideline or a drug label — the reference. Where the evidence is genuinely contested, it says so rather than guessing.
The correct answer is never asserted at you. You get the mechanism, step by step, with the source behind each step — and each wrong option explained, with the reason it is wrong rather than just a cross.
Each question records the exact nodes it tests, so a wrong answer points at specific mechanisms — and at the neighbouring ones you probably have not met yet.
Questions are re-derived from the underlying medical knowledge base rather than written once and left. When that knowledge grows, questions you have already seen get better.
A full paper, in the published format, assembled before you start — so nothing loads while you are sitting it. No feedback until a block is submitted, and a submitted block does not reopen. Your answers are saved as you go, so a refresh or a lost connection costs you nothing.
Questions are drawn in the proportions the exam publishes. Where an exam publishes no weighting, areas are weighted equally and it says so.
Pick a specialty, or type any subject — a whole field or one narrow corner of it.
Pick a subject — a whole field or one narrow corner of it. The system alternates teaching and questioning, and which one comes next is decided by what it currently believes about you: unknown means ask, consistently-wrong means teach, and straight after a card it always asks, because teaching tells us nothing about you.
Colour is how well you know each area. A dashed ring means the estimate comes from neighbouring topics rather than from questions you have actually answered — a suspicion, not a result.