Sense
Most play candidates should never become plays. The sensing discipline is noticing repetition without rushing to codify it.
Play: notice the third time
When: you have run the same move in roughly the same situation three times and it worked each time, or you catch yourself explaining the same approach to a third person.
The move: write a one-line note naming the situation and the move. Nothing more. Park it in a single list of play candidates.
Signals: the candidate list grows faster than the playbook. Most candidates never graduate, and that is correct.
When not to: after one success. One success is an anecdote.
Antipattern: premature codification
Writing the play after running the move once. The play gets written from hope rather than evidence, its entry conditions are guessed, and its signals are aspirations. Premature codification is how playbooks fill with plays that never transfer, and it is the most tempting failure for a new collection because early on the pressure is to have volume.
The counter-discipline is a two-list system. Candidates live in one place, plays in another, and the bar between them is explicit: multiple uses, observed or run first-hand, with at least one instance you can narrate in detail. Three plays that transfer beat fifteen that do not.
Hypothesis: three uses is the right default threshold. I have not tested whether the number should flex with the risk of the play, though I suspect it should.
Licensed under CC BY-SA 4.0.