Tool 32 of 32 · a few minutes
Ledger Read-Back
The due forecasts read out next to what actually happened, because the gaps are the fastest way to learn what the plan got right.
Inside each Checkpoint Review, after the re-measure, read by a neutral person, never by the forecaster.
What it is for
To read the due Forecast Ledger entries next to what actually happened, so the admissibility the ledger produced actually transfers, and so the gaps teach what the plan got right.
The failure it catches
The Forecast Ledger exists and never gets read, so the admissibility it mechanically produced never transfers. Or it gets read by the forecaster as vindication, which spends the credibility instead of banking it.
Why it works
A neutral reader, forecast next to actual, no commentary attached to names, turns hit rate into arithmetic the room can use. Kept entries and broken entries get the same calm question, so nobody has to say I told you so, which is the sentence that kills the book.
How to run it
- At each checkpoint, a neutral person, never the forecaster, reads the due entries: forecast next to actual, kept, broken, or open, no commentary attached to names.
- Kept entries get one question to the room: what does this hit rate say about whose conditionals we weight going forward.
- Broken entries get the symmetric question, asked with the same neutrality.
- Before the room disperses, log new entries arising from the checkpoint's adjustments: if this, then that, by when.
- At close, the ledger hands over as standing practice with the read-back rule attached: a neutral reader, no names on broken entries.
How to read the result
The running kept-rate, updated in the open. After two or three kept entries the weighting conversation is arithmetic, not argument. That transition happening, or not, is the reading. Admissibility transfers to the people whose conditionals keep coming true, mechanically, without anyone spending credibility to claim it.
How to run it well
- The neutral reader and the no-names rule are both load-bearing.
- Ask the kept question and the broken question with the same temperature.
Where it breaks
- Read by an insider it becomes I told you so and burns.
- Read with names attached to broken forecasts it becomes a blame tool and the entries stop coming.
What would prove it wrong
A strong kept-rate, read back neutrally at three checkpoints, that still shifts no decision weight would mean read-back does not transfer admissibility here.
What we watch
Whether a strong kept-rate, read neutrally, actually changes whose conditionals get weighted, or whether the room nods and then decides the same way it always did.
Ask yourself
- When a forecast came due, who read it back, the person who wrote it or someone neutral?
- Was the forecast sitting next to what actually happened, kept or broken, with no name attached to the miss?
- After two or three kept entries, did anyone's written conditionals actually get more weight on the next call?
- Before the last review ended, did new if-then-by-when entries go into the ledger?
- If someone read a kept entry as I told you so, did entries keep coming afterward?
Fed by: Forecast Ledger
Feeds: Checkpoint Review, Forecast Ledger