Most candidates prepare for a language test the way they prepare for an exam at school: they work on everything, a bit, and hope the overall picture is good enough.
The ICAO rating scale does not work that way. Six descriptors are scored separately - pronunciation, structure, vocabulary, fluency, comprehension and interactions - and your final rating is the lowest of the six. Not the average. The lowest.
The practical effect is brutal and useful at the same time. A pilot with excellent vocabulary, easy fluency and sharp comprehension can still be held at Level 4 for years because of one thing: a vowel that collapses under stress, or a habit of answering before checking, or a grammar structure that falls apart when the situation gets hypothetical.
It is brutal because all that general study did not touch the thing that was actually costing you the level. It is useful because it means the work is far smaller than you think. Find the weak descriptor, work on that, and the rating moves.
This is why every course I run starts with a diagnostic rather than a textbook. Twenty minutes of conversation and one non-routine scenario is usually enough to see which of the six is dragging the others down. Everything after that is aimed at the same target.
