goal-backward-planning

Outcome-based planning reverses from a verifiable end state to required preconditions. You define the observable goal, derive the truths that must hold, and generate artifacts and dependency-aware task waves to reach it.

Backward planning structures a roadmap by translating an expected outcome into verifiable preconditions, required artifacts, and parallelized task waves. This ensures implementation follows from testable dependencies rather than arbitrary activity lists.

You group tasks into wave-based parallelization cycles ordered by real dependencies. This sequencing ensures parallel work is safe and efficient by preventing execution until required observable truths and artifacts are verified.

Yes, it applies to software work planning and roadmap breakdown. It is suited for complex changes like secure authentication updates where you must define token behaviors and required tests before deciding implementation details.

The best way is using an explicit checkpoint protocol that inserts human verification and irreversible decision gates. This prevents moving forward without validation at critical decision points in the delivery roadmap.

You select discovery levels to determine the minimum necessary investigation depth. This reduces wasted context by ensuring you only gather the required observable truths and artifacts needed for the specific goal.