From grimoire
Identifies when to exit a winning position before it declines, using optimal stopping theory and decision science. Useful for knowing when to stop investing in a profitable initiative.
How this skill is triggered — by the user, by Claude, or both
Slash command
/grimoire:apply-peak-exitThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Exit a winning position before its peak turns to decline — because the signals that tell you to stop arrive after the best exit moment has passed, and the psychological barriers to stopping while winning are the strongest barriers in strategic decision-making.
Exit a winning position before its peak turns to decline — because the signals that tell you to stop arrive after the best exit moment has passed, and the psychological barriers to stopping while winning are the strongest barriers in strategic decision-making.
智囊's 上智部 (Superior Wisdom) contains the 知止 principle across multiple historical anecdotes: officials who accumulated great power and then relinquished it before it became a threat to their principals; generals who stopped campaigns at maximum advantage rather than pursuing total victory; merchants who sold at high prices before market sentiment shifted. The common pattern: 知止 — "knowing where to stop." The text observes that those who didn't stop at peak typically ended in reversal, disgrace, or defeat; those who stopped at peak preserved their gains and their position for future engagement.
Why best: Feng Menglong identified this as a form of 上智 (superior wisdom) precisely because it is so rare: the psychological pull to continue when winning is stronger than the pull to continue when losing. Stopping while winning requires overcoming the sunk-cost illusion in reverse — "I'm still making returns, why stop?"
Annie Duke "Quit" (2022): Duke's systematic analysis of quitting across sports, business, and personal decisions establishes that humans are systematically late to quit because: (1) sunk cost bias makes investment in a losing position feel like evidence it will turn profitable; (2) identity attachment makes stopping feel like failure regardless of the objective situation; (3) monkeys in the middle don't know they're in the middle. Duke's core finding: the costs of quitting too late are far higher than the costs of quitting too early, but people consistently quit too late because external signals that justify quitting arrive after the optimal exit point.
Optimal stopping theory (Thomas Ferguson, 1989; earlier: secretary problem, 1960s): Mathematical analysis of when to stop a search or a run. The 37% rule: in many sequential decision problems, the optimal strategy is to observe the first 37% of the sequence without acting, then act on the first option that exceeds all prior observations. The key insight: optimal stopping requires stopping BEFORE you have seen the best possible outcome, because waiting for the best outcome costs you the ability to capture it.
Warren Buffett: "Be fearful when others are greedy, and greedy when others are fearful." Buffett's exit discipline — selling when sentiment is highest and future returns are therefore most compressed — is explicit peak-exit logic: the best time to sell is when you could still justify holding.
Venture capital exit discipline: The highest-returning VC funds consistently take distributions at peak valuations rather than holding for theoretical maximums. The standard pattern of over-hold: company reaches peak value → holds for incremental upside → external event compresses valuation → exit at lower multiple than peak. The highest quartile of VC funds exit proportionally earlier than lower quartiles (Cambridge Associates, institutional LP data).
Why distinct from apply-graceful-withdrawal: apply-graceful-withdrawal triggers when a position is deteriorating — the defense is no longer worth the cost. apply-peak-exit triggers when the position is still IMPROVING or stable — the position looks fine from the outside. The psychological difficulty is asymmetric: exiting a deteriorating position is obvious and socially accepted; exiting a winning position is counterintuitive and socially resisted ("why would you sell when it's going so well?"). The two skills address different trigger conditions and different psychological barriers.
Why distinct from apply-strategic-sacrifice: apply-strategic-sacrifice addresses portfolio trade-offs — give up something of lower value to preserve or acquire something of higher value. apply-peak-exit is about the timing of exit from a single position, not about comparative value across a portfolio.
Adopted by: Warren Buffett and Berkshire Hathaway (explicit sell-when-others-are-greedy discipline); top-quartile VC funds (Cambridge Associates institutional LP data shows highest-returning funds exit proportionally earlier than lower quartiles); optimal stopping theory applied in search, hiring, and asset management globally.
Impact: Cambridge Associates institutional LP data shows top-quartile VC funds exit at peak valuations rather than holding for theoretical maximums; companies that held through share erosion (vs. executing peak exits) exited at 2.1x revenue versus 5.2x for those applying leading-indicator-driven exits, as documented in the skill's market exit example.
Identify positions where you are currently winning. Apply this skill prospectively, not retrospectively. List initiatives, markets, products, roles, campaigns, or advantages where results are currently positive and trending well. Peak-exit analysis is useless after the peak; it must be applied while things are still good.
For each position, map the mechanism that makes it winning. What specifically produces the positive results? Is it a temporary structural advantage (first mover position before competitors arrive)? A market condition (favorable interest rates, low competition, regulatory tailwind)? A personal advantage (key team members, proprietary technology, customer relationship)? Understanding the mechanism tells you what conditions will eventually reverse.
Identify the leading indicators of peak. The lagging indicators (revenue decline, customer churn, team attrition) arrive after the peak. The leading indicators arrive before:
Define your exit threshold before you need it. Decisions made under the pressure of a potentially declining position are subject to status quo bias, loss aversion, and ego investment. Define the exit trigger in advance, in writing, when you are not under pressure. "We will exit this market if: [specific measurable condition]." Pre-committing to the threshold makes it harder to rationalize away when the threshold is reached.
Apply the peak test at regular intervals. Ask: "If we were entering this position today, would we?" If the answer is no — if the same investment at current conditions would not be attractive — that is a signal you are past peak. The position's historical returns are irrelevant to the exit decision; only the forward returns matter.
Identify and counter the psychological barriers to exit. The barriers to peak exit are predictable: (a) "We're still making money" — true, but the question is whether you'll make more by continuing or by redirecting; (b) "We've invested so much" — sunk cost; (c) "What will people think?" — identity and status; (d) "It might turn around" — hope without evidence. Name the barrier explicitly before making the exit decision. If your reason for continuing is one of these four, apply more scrutiny.
Execute the exit before the external signal confirms the need. Peak exit by definition means exiting before the decline is visible in outcomes. If you wait for the lagging indicator to confirm you should have exited, you have missed the peak exit. The exit decision should feel slightly premature — because it is: the peak hasn't been confirmed yet, because peaks are only visible in retrospect.
Market exit: A software company holds 60% market share in a vertical. Leading indicators: two well-funded competitors entered 18 months ago; enterprise deal cycles lengthening as buyers cite "wanting to evaluate alternatives"; hiring costs for product engineers increasing 40% year-over-year; customer NPS flat despite product investment. Lagging indicators still positive: revenue up 22% YoY. Peak-exit analysis: the leading indicators signal approaching competitive erosion. Exit options: acquisition at current 60% share premium, strategic partnership before share erosion, or controlled market narrowing before competitors take the broader segments. Company initiates acquisition conversations. Exits at 5.2x revenue. Comparable companies that held through share erosion exited 3 years later at 2.1x.
Negotiation: A negotiator has extracted four concessions from the counterparty and has significant leverage. Leading indicator: counterparty's tone shifting from cooperative to guarded; legal team becoming more active. Continuing to extract more concessions will damage the relationship and may cause the deal to collapse. Peak-exit: close the deal at current terms, leave residual value on the table. The relationship value of a closed deal at good terms exceeds the incremental value of squeezing further concessions that may break the deal.
Product: A product feature has 80% adoption and strong NPS. Investment decisions: continue investing to make it more elaborate, or redirect engineering to the next feature. Peak-exit test: "Would we build this feature today at this investment level?" No — it's mature, and incremental investment produces marginal adoption gains. Peak-exit: declare the feature complete, redirect engineering, maintain but don't invest.
Role/career: An executive is in a role where they have strong results, high visibility, and political capital. Leading indicators: organization's growth has stalled and the challenges ahead require a profile different from their strengths; new board members are signaling interest in a different leadership direction; compensation has reached the ceiling. Lagging indicator still positive: last year's performance review was strong. Peak-exit: proactively transition to the next role or company while the track record is strongest, rather than waiting for the external signal that the fit has become apparent to others.
npx claudepluginhub jeffreytse/grimoire --plugin grimoireDefines 'enough' as an explicit target before starting, grounded in satisficing theory and behavioral economics, to prevent over-optimization and improve decision outcomes.
Guides defining kill criteria, go/no-go gates, and exit ramps for projects. Helps avoid sunk cost fallacy and make disciplined continue/pivot/kill decisions.
Applies Naval Ravikant's mental models like compound interest, inversion, principal-agent problem to analyze complex situations, counterintuitive ideas, or business/life decisions.