You took four mocks. The scores were close together. The natural conclusion is that you have hit your level, and it is usually wrong — a plateau is more often a sign that the mocks are being taken but not used.

You are measuring rather than changing

A mock is a measurement. Measurement does not cause improvement, any more than weighing yourself causes weight loss. If nothing about your revision changed between mock three and mock four, there was no mechanism by which the score could move.

The question after every mock is not "what did I get?" but "what am I doing differently this week because of it?"

The same errors are recurring and you have not noticed

Without an error log, each mistake feels fresh. With one, you discover that four of your last twenty errors were the same misconception about equilibrium — which is a single afternoon's work, not a general weakness.

People consistently overestimate how varied their mistakes are.

You are revising what is comfortable

Revision naturally drifts towards topics that feel good to revise, which are the ones you already know. It is pleasant and it changes nothing.

The chapters you avoid are avoided for a reason, and that reason is where the marks are.

Time, not knowledge, is the binding constraint

Check whether your unattempted questions are ones you could not do, or ones you never reached. If the latter, no amount of further revision helps — the fix is pacing, and pacing is a separate skill practised separately.

You are guessing more than you think

Track it. If a meaningful share of your correct answers came from narrowing and guessing, your real score is lower than the number, and it will regress the moment the paper is slightly harder.

That is not pessimism; it is the difference between a score you own and one you were lent.

What to do this week

  1. Take the last three mocks and list every wrong answer in one place.
  2. Group them by topic, not by paper.
  3. Any group with three or more entries is this week's revision. Ignore everything else.
  4. Take the next mock and check whether those groups shrank.

That loop is the mechanism. Without it the mocks are just a record of the plateau.