Three horizons
The plan is disciplined by the horizon above it. Without a +10 layer, five-year targets drift toward whatever is achievable in five years — the filter that removes exactly the goals worth having.
2031
The plan · what must be true
Revised only at the mid-term.
2036
Long-range objectives · +10
Fixed by 15 Nov 2026. Revised only at the terminal review.
2051
The line · +25
Due 30 Sep 2026. If it never changes a decision, it was decorative.
Goal catalogue
A menu, not a checklist. Thresholds for domains currently below the line are in scope now. Four to six ambitious goals across five years. Exactly three lofty ones, across uncorrelated failure modes. Adopting a goal creates its anticipatory target and a binding stub you then have to quantify.
The plan
Ceilings: eight active binding targets, three lofty goals, four domains at major emphasis or above. These are limits, not objectives.
The rule that does the work: never set a binding target on another person's behaviour. Friendship, love, recognition and audience are all consumption-share problems — they matter most, cannot be commanded, and the only honest move is to command the exposure and initiation that make them likely.
Mechanism & rhythm
Left: converts an outcome into the input rate that produces it. Right: the one cross-cutting number — standing weekly commitments involving other people that would be noticed if missed.
—
Standing weekly commitments
0
target band 4–6
Emphasis
Click to cycle 0–3. Constraints are shown, not enforced — the point is to see the trade-off, not to be prevented from making it. With twelve domains, most cells should be 0 or 1.
0 · dormant, deliberately1 · maintained2 · major3 · defining
Gating decisions
A question without a date isn't open, it's avoided. Any decision gating three or more domains is forced in Year 1 — the carrying cost of deferral compounds; the cost of choosing wrong is bounded.
Reviews
Ninety minutes, four times a year, hard stop. A missed target cannot be left unclassified. Retire at least one target at every review — a plan that only grows gets abandoned whole rather than pruned.
A — delete the target, one sentence on why it looked right.
B — recompute the input rate. Raise it or extend the deadline. Not both.
C — log the assumption that broke, then reset.
D — change nothing about the target. Reduce the number of active binding targets instead.