1.3 Scaled Scoring, Attempt Limits and Best-of-Five Technique

Why There Is No Percentage Pass Mark

Candidates routinely ask "what percentage do I need?" and the honest answer is that the question does not apply. Since the 2008/3 diet, MRCP(UK) Part 1 results have been calculated by equating, a statistical process based on Item Response Theory (IRT). Because no two diets contain the same questions, one paper may be marginally harder than another. Equating calculates the difficulty of each individual item and converts your raw score into an overall scaled score so that identical ability produces an identical result whichever diet you sat.

Reading Your Score Report

The change of standard matters practically. The scaled pass mark was reset to 450 from the 2026/1 diet following the MRCP(UK) Part 1 Standard Setting Group's October 2025 exercise. Any revision material, forum post or older question bank that still quotes 540 as the target is out of date, and a candidate calibrating against 540 is working to a standard that is no longer in force.

Two further implications follow from equating:

• You cannot compute your result from a raw percentage. The Federation's own worked illustration for its equated specialty exams shows that the same equated score can correspond to 61% correct in one paper and 59% in a harder one. Treat mock percentages as a relative trend, not an absolute prediction.
• Your performance is not graded against the other candidates in your sitting. Equating anchors to item difficulty, not to a fixed quota of passes, so a strong cohort does not raise the bar for you.

Attempt Limits

The General Medical Council requires the medical royal colleges to cap examination attempts, and has specified a maximum of six. The Federation therefore allows:

1. A maximum of six attempts at each part of the MRCP(UK) Diploma — Part 1, Part 2 Written and PACES are counted separately.
2. One exceptional additional (seventh) attempt, applied for on the basis of additional educational attainment, and no more than one.
3. A seventh-attempt application signed by your educational supervisor and training programme director confirming that improvement makes a pass highly likely, submitted at least two weeks before the relevant application period opens, and decided by the MRCP(UK) Regulatory Sub-Group, whose decision is final.

Withdrawals do not count as attempts. This is a genuinely useful protection: the Federation states that any candidate who starts an examination is declaring themselves fit to sit it, and you have the right to withdraw at any point up until the exam starts without penalty. If you are acutely unwell on the morning of the exam, withdrawing preserves an attempt; sitting and failing consumes one. Requests for additional attempts based on exceptional circumstances are not accepted — the correct route if performance was affected is to inform the Federation promptly and appeal after results.

Best-of-Five Technique

The format is one correct answer and four distractors that are "closely related to the preferred option but less correct". Technique therefore has to be about ranking, not spotting.

• Read the last line first. "Which is the most likely diagnosis?", "What is the most appropriate next investigation?" and "What is the single most appropriate immediate management?" send you through the same stem in completely different directions.
• Commit to an answer before reading the options. Generating your own answer from the stem protects against distractors engineered to look attractive.
• Watch the qualifier. Most likely asks for probability; most appropriate next step asks for sequence and often rewards the cheap, safe, immediate action over the definitive one; definitive asks for the endpoint.
• Use demographic and geographic anchors. Ethnicity, travel history, occupation and drug history are rarely decorative in a well-written best-of-five stem.
• Do not change a considered answer without a reason you can articulate. Second-guessing under time pressure converts more right answers to wrong than the reverse.

Pacing Arithmetic

At 180 minutes for 100 items you have 1.8 minutes (108 seconds) per question. A workable model is:

Aiming to finish at 170 minutes leaves ten minutes to revisit flagged items and, critically, to ensure no item is left blank. Rehearse this in full 100-question blocks — pacing is a trained skill, and the six-hour, two-paper structure means fatigue management in paper 2 is as much a determinant of your scaled score as knowledge is.