reframe

REFRAME — Representation Multiplier

You are solving this problem in one representation — one mental model, one framing. That frame feels natural. It is also a cognitive prison: the search dynamics change with representation. A dead end in one frame is often trivially solvable in another.

Hard problems stay hard because the solver never changes the lens.

---

The Failure Mode You Must Recognize

You are about to solve the problem the same way for the third time — using the same abstractions, the same vocabulary, the same decomposition. Each attempt produces slightly different results but no breakthrough. This is single-frame reasoning: the model is optimizing harder within one representation instead of switching to a representation where the answer is easier to find.

---

The Protocol

Step 1 — NAME THE CURRENT FRAME

Write down how you are currently framing this problem:

CURRENT FRAME
────────────────────────────────────────
Problem:                [one sentence]
Current representation: [how you're modeling it — e.g., as a data flow,
                         as a user story, as a code structure]
What this frame reveals:[what's easy to see in this frame]
What this frame hides:  [what's hard to see — be specific, not "some things"]
────────────────────────────────────────

Artifact: The current frame analysis. You cannot escape a frame you haven't named.

Step 2 — CHOOSE THREE DIFFERENT LENSES

Select three from this set that are DIFFERENT from your current frame. No near-duplicates:

Lens	What It Shows	Use When
Constraint graph	Nodes = decisions, edges = constraints. Reveals bottlenecks and dependencies.	You can't figure out what's blocking the solution
State machine	States the system can be in, valid transitions. Reveals illegal states and missing transitions.	The problem involves stateful behavior or workflows
Cost function	What you're minimizing/maximizing. Penalty surface shape.	You need to make an optimization tradeoff
Queue & flow	What enters, what exits, where things back up.	Performance, throughput, or capacity problems
Risk matrix	What can go wrong × how bad × how likely.	You need to prioritize what to worry about
User journey	The experience as a story: trigger → action → outcome → feeling.	The problem is about UX or human behavior
Data model	Entities, relationships, invariants (what must ALWAYS be true).	The problem is about correctness or consistency
Decision tree	Branching choices with outcomes at leaf nodes.	The problem involves sequential decisions