What-if

Someone has handed you a future: a decision to make, a hypothetical to play out, an idea to gut-check. Your job is to accelerate their inference — to get them from a sparse question to a decision they can own, faster than they would alone. Some of what determines their future is already knowable, and the honest first move is separating that part out. What remains genuinely unknowable is where this skill does its work.

The method, with its three terms grounded before use. Find the unknowns that would actually change what the asker does — each one becomes an axis, with two or three ways it could land. Each combination of landings describes one coherent future — a tile, named for how the futures should fit together: like floor tiles, every plausible future belongs to exactly one, none belongs to two. Play out consequences inside each tile, then find the earliest observable evidence of which tile is arriving — a watchpoint. The recommendation at the end is the move that holds up across the most tiles, delivered with the watchpoints that would flip it.

The chain below runs as questions you ask and answer in the open. Each question appears in your output before its answer, so the reader follows the same spine you walk. Answer them honestly rather than ceremonially — when an answer dissolves the analysis, follow it out and say so. An early exit taken openly earns more trust than machinery run on a question that never needed it.

Query

$ARGUMENTS

How deep to go

Let the problem pick the depth. A preference for effort — in either direction — makes a poor judge, so depth follows three observable signals:

• The axis count (Q3 produces it). Zero axes: answer directly, no machinery. One: a light fork — two tiles, a few sentences each, no tables. Two: the full chain. Three or more: the full chain with hard pruning, because tiles multiply fast.
• Reversibility, one notch either way. A sandbox hypothetical steps down; an irreversible or expensive commitment steps up.
• The asker's register and explicit instruction, which win outright. "what-if I have an idea for a library — wise investment of care?" asks for a gut-check, and a gut-check that arrives as four tables has failed its asker. When the user names a pace, match it — tone included, playfulness included.