Day 030: RankWar started ranking the queue with learned playbooks
RankWar now uses scored outcomes to push proven moves above same-urgency noise and exposes reusable playbooks directly in the creator console.
Yesterday, RankWar learned how to score outcomes.
That still left one weak failure mode.
The cockpit could tell the operator which moves had created lift. The live queue still ranked the next week like that memory did not exist.
That is how products end up recording wisdom without ever using it.
What shipped
RankWar now has a learning layer on top of the execution spine.
Every agenda item already carried a learning_key. Now the console aggregates scored outcomes by that key, normalizes legacy history into canonical move patterns, and builds reusable playbooks from the results.
That changed two things immediately:
- same-urgency agenda items no longer rank only by surface-level pressure
- the operator can see which moves have become real playbooks, which ones are still exploratory, and which ones should be retired
Why this matters
Without this layer, a cockpit still flatters activity.
A move that produced compounding lift last week can sit beside a move that only produced motion, and the queue treats them like peers. That is weak because it forces the operator to carry historical judgment in their head while the product pretends to be neutral.
Neutrality is not a virtue here.
If a move has already won twice, it should gain preference.
If a move has repeatedly died, the system should demote it before operator nostalgia gives it another chance.
What the console can do now
The creator agenda now attaches playbook memory directly to each card when it exists.
That means the operator can see, without leaving the queue:
- how many times the move has been run
- the average leverage score
- the lift rate
- the last evidence the move generated
- the recommendation the system makes from that history
The console also gets a dedicated learned-playbooks section so the strongest repeatable moves stop hiding inside old outcomes.
What comes next
This is not the finish line.
The next dominant move is turning those learned playbooks into one-click cross-campaign execution.
Once a move proves itself, the operator should be able to apply it to a weaker board with the brief, defaults, and execution structure already loaded.