What a practice score can and cannot tell you

Many learners treat the latest practice score as a verdict: 64 on Monday means a good week, and 57 on Thursday means a disaster. Neither number deserves that much feeling. One score is a rough reading from one sitting, and what counts is what several of them show together.

What the scores on this site are

The scores and feedback on this site are AI-estimated. They are there to guide your practice. They are not official results, and nobody can promise a particular result in the real test. The real test is scored by Pearson's own automated system, which this site cannot see inside, so a number here should never be read as a forecast of a number there.

What a series of scores can show

Look for patterns across attempts, not for the best number or the worst. This is an illustrative example with made-up numbers: five attempts over three weeks.

• Overall estimate: 58, 60, 57, 59, 61
• Speaking: 55, 60, 52, 58, 61
• Writing: 66, 68, 65, 67, 69
• Reading: 60, 64, 58, 62, 63
• Listening: 50, 49, 51, 48, 50

Two things stand out. First, the overall figure bounces by a few points from one attempt to the next, and the line may be edging up, though five attempts are too few to be sure. "About 59, perhaps rising" is a fairer summary than 61 or 57. Second, Listening sits near 50 every time, the lowest line on the sheet and the flattest. A flat, low line is the signature of a consistently weak skill, and it tells you where next week's time should go.

Notice what the numbers do not say. They do not say the Writing result is "safe", only that it is steady. They cannot explain why Speaking dropped eight points between the second and third attempts. That needs the written feedback and a look at which tasks were in each set.

What they cannot show, and why the real test differs

• Your real score. A practice result comes from a different scoring system, and the two should not be expected to agree exactly.
• Different content. The real test uses different questions. A set of easy passages, or a hard set of images, moves a result without any change in you.
• Familiar tasks. If you have seen a question before, a better result may owe more to memory than to skill.
• Nerves. A test centre, a countdown and a lot riding on the outcome change how people speak and type.
• Equipment. Your headset and room at home are not the equipment at the test. Background noise can cost you in Speaking and Listening.
• Stamina. The real sitting lasts roughly two hours. Concentrating that long is a skill of its own, and a short practice session does not test it.

If you need a particular score for study or work, requirements differ and change, so check the official source for the figure you need. A practice score cannot answer that question.

How to use your scores

1. Keep a simple sheet: date, overall, the four skills, and one line of notes.
2. Take at least three attempts under similar conditions before drawing any conclusion: the same headset, a similar time of day, no pausing.
3. Compare the skill breakdown, not only the overall number. Ask which skill is lowest each time.
4. Ignore one-off swings. Take the middle value of your last three results, not the best or the worst, as your working figure.
5. Change one thing at a time, such as a new listening routine, and see whether the line moves over the next two or three attempts.

In the notes column, write down sleep, noise and mood. When a score looks odd, that line often explains it.

For advice on when to take full tests and what to do afterwards, read using mock tests well. For the written feedback, see reading AI feedback on your practice answers sensibly.

Tonight, put your last few practice results side by side and find the one skill that never moves. Work on it for a week, then take another attempt in the practice tests on this site and see whether that line has shifted.