Tool 20 of 32 · a few minutes
Forecast Ledger
A dated record of who expected what, so the company stops arguing about the past and starts weighing written accuracy.
Installed on decision day as a standing practice, then read back at every checkpoint.
What it is for
To stop the company arguing about the past by writing down what people expect from the future, in a form that can be wrong on a date.
The failure it catches
Perfect prediction, zero authority transfer. The most accurate people in the building often carry the least weight. Being right after the fact wins arguments and changes nothing, especially when the insider reads their own vindicated forecast back.
Why it works
Dated, specific conditionals turn accuracy into a hit rate. After a handful of entries come due, weight moves toward the people whose writing keeps coming true, without anyone having to fight for it.
How to run it
- Start a shared, dated record. Every entry takes one shape: if this decision or condition, then this specific consequence, by this date.
- Both the person forecasting and the person deciding acknowledge the entry when it is made.
- When the date arrives, log the outcome next to it. Vague dread does not qualify. This will bite us eventually is not an entry. If we ship without the migration, support tickets double within six weeks is.
- Have a neutral holder keep the ledger, especially at the start. Entries land differently when the person who was right is not the one holding the book.
- New entries arising from later adjustments go in before the room disperses.
How to read the result
After a handful of entries have come due, the company stops arguing about whose opinion should count and starts noticing whose written conditionals keep coming true. That transition happening, or not, is the reading. It changes the second disaster, not the first.
How to run it well
- Entries must be falsifiable and time-bound or the ledger becomes a grievance file.
- During a diagnosis, a neutral party holding the ledger is the active ingredient.
Where it breaks
- If anyone reads it back as an I-told-you-so, the ledger dies. Its whole value is that nobody has to say that sentence anymore.
- Reading it with names attached to broken forecasts makes it a blame tool, and the entries stop coming.
What would prove it wrong
A ledger with a strong kept-rate that still changes no decisions after months would mean written vindication does not buy admissibility here.
What we watch
Whether a running kept-rate actually shifts who gets believed on the next call, or whether the book fills up and nothing moves.
Ask yourself
- When was the last time someone wrote down what they expected to happen, with a date, before it happened?
- Whose written predictions keep coming true, and does that person have any extra weight in the next decision?
- Are entries specific enough to be wrong by a date, or are they vague dread?
- Who holds the record, and would an insider reading it back as vindication burn it?
Fed by: Padding Probe, Ledger Read-Back
Feeds: Ledger Read-Back