This is the part that makes the method unusual, and it is the part that costs something to keep.
The prediction is recorded. Every instruction goes into a standing register with its confidence range and the date it will be checked. It is written down before the outcome is known, which is the only time a prediction means anything.
The check happens at twelve months. The instruction is compared against what actually occurred. Not against a revised memory of what was intended.
The result is published either way. A method that only reports its successes is a marketing programme with a scoring system attached.
Almost no strategy framework commits in advance to being marked. Advice is given, the engagement ends, and nobody returns to check. That is not dishonesty. It is simply what happens when nobody has an obligation to look.
Each volume carries a calibration target: the rate the method is built to hit. Those targets came from development work, and they are stated with the number of cases behind them, which is usually somewhere between fourteen and forty-two. That is a small number and it is written as a small number.
They are not observed results from client engagements, and they should not be read as a prediction of performance in any particular situation. The register replaces them with observed figures as it fills.