The first question in any study plan is how many practice tests to do. The short answer: there is no universal number; what matters is how much each paper teaches you and whether it is still unseen.
Why repeating a paper flatters you
The second time through a Reading paper you are not reading, you are recalling. The score rises while the skill does not. In Listening it is starker still: once you remember an answer, that recording has lost its measuring value permanently.
So a completed paper is for error analysis, not for measurement. A four-test book used properly gives you four measurements, not twelve.
Count the things worth counting
Instead of counting papers, track three numbers over time:
- Correct answers out of 40 per skill, dated.
- Distribution of errors by section and by question type.
- Repeat errors: how many consecutive papers show the same failure.
The third number matters most. A failure appearing in five consecutive papers is not bad luck — it is an unrepaired gap, and more papers will not repair it.
When you run out of unseen material
Practice banks are finite; preparation time is not. This is why serious platforms assemble papers from components: existing reading passages, writing tasks and listening sections are recombined into new papers that follow the real structure.
On LinguaGrade each generated paper draws its components from several sources, with a reuse ceiling per component so that one reading passage does not reappear too often across the papers you sit. That preserves the essential condition of a measurement: you have not seen this content before.
The condition attached is structural validity. A generated paper still needs 40 questions in Listening and Reading, two Writing tasks, three Speaking parts, and an answer key for every objective item. A carelessly assembled paper produces a number less trustworthy than a repeat.
A study rhythm that works
- Weekly: one full four-skill paper under real timing.
- After each paper: two hours of analysis, longer than the paper took.
- Midweek: drill the question type you failed most last week, without sitting a full paper.
- Monthly: review the three numbers above and adjust.
Frequently asked questions
Will forty papers guarantee a higher band?
No. Paper count does not correlate with band improvement without analysis between attempts.
Are generated papers like the real test?
Structurally yes, when components meet the standard. They are not statistically calibrated on thousands of candidates the way live papers are, so use them for practice and trend, not as a certified band.
How close to test day should I keep practising?
Keep a steady rhythm rather than cramming. In the final week, reduce paper volume and increase review of known errors.