Tool · questions

Ten questions before deploying any capital

1. What is the holding horizon?

   Money needed in two years and money free for twenty cannot carry the same risk, because horizon decides how much interim decline is survivable without being forced to sell. The asset-menu chapter maps which instruments are even eligible at which horizon. An answer of "indefinitely" is usually a sign the question has not been thought about, not a sign of patience.

2. What is the maximum acceptable permanent loss?

   Not a temporary dip — a permanent impairment, written as a percentage of the position and of the whole portfolio. Losses compound asymmetrically (a 50 percent fall needs a 100 percent rise to recover), which is why the average-returns chapter treats the left tail as the number that matters. The figure is written before entry because tolerance measured during a fall is always lower than tolerance imagined before it.

3. What macro regime is in force right now?

   Assets behave differently as growth and inflation shift — buying is choosing what to wear, and the regime is the weather, per the regimes-as-weather chapter. If the current combination of rising or falling growth and rising or falling inflation cannot be named, the purchase is a guess about conditions the buyer has not read. "Unclear" is an acceptable answer; it simply argues for smaller size.

4. What is the portfolio's current stance?

   Risk-leaning, defensive, or neutral — with neutral as the honest, disciplined default when evidence is mixed, because conviction is a resource that must be earned. The discipline-layer chapter develops this three-state idea. A purchase should be consistent with the declared stance; a single exciting idea is not a reason to override it.

5. Is the reserve covered first?

   A reserve floor is capital that is never deployed regardless of how good the signal looks, because the best opportunities appear exactly when everyone else is forced to sell. The floor is a wall, not a dial. If the purchase can only be funded by reaching into the reserve, the answer is already no — no further analysis required.

6. Is the instrument appropriate to the stance?

   Direction and vehicle are separate decisions: a cautious view expressed through an aggressive instrument — borrowed money built in, upside sold away, or a deep lock-up — is a contradiction in structure. The liquidity-ladder chapter makes the vehicle question concrete: a T-bill's rung, a savings bond's rung, and a decades-locked rung are different assets even at the same headline return. Match the rung to the stance, not to the brochure.

7. Is the position crowded or contrarian?

   Prices move on who is forced to act next, not on who is right — the positioning-and-flows chapter's core. A crowded position exits violently, because everyone heads for the same door at the same time. Knowing which side of the room the crowd stands on is part of the price being paid, whether or not the thesis is correct.

8. What is the worst realistic scenario?

   This is the CVaR question from the VaR-to-CVaR chapter: not the everyday bad day, but the average outcome inside the worst slice — when it goes bad, how bad is bad on average? VaR marks the door to the bad region; CVaR measures the room behind it. If that outcome is unsurvivable at the intended size, the size is wrong even if the idea is right.

9. What is the exit if the thesis is wrong?

   Written in advance: what evidence would falsify the idea, at what point the position is closed, and at what cost — including any notice periods or penalties. A position without a pre-committed wrongness condition is not a thesis; it is a hope with a ticker. Lock-ups make this question sharper, not optional.

10. Does the expected return justify the risk AND the complexity?

    Return is only half a number; the other half is what was risked to get it. Complexity is a separate, additional cost — more moving parts, more counterparties, more silent failure modes — so the claim must beat a boring alternative after fees and after risk, at a size the Kelly chapter's fractional logic would call survivable. If the honest answer is "only if everything goes right," the answer is no.

The list exists because the expensive mistakes are rarely analytical — they are procedural. The investor knew the product was complex, suspected the position was crowded, never wrote down an exit, and deployed anyway. Ten questions, answered in writing before commitment, convert that failure mode from likely to rare.

How to use it

Answer in writing, dated, before any commitment — memory rewrites itself once outcomes are known, and an undated answer cannot be audited later. The first five questions are about the investor and the portfolio; the second five are about the position. That order is deliberate: a position cannot be judged until the horizon, loss tolerance, regime, stance, and reserve it must live inside are stated.

A blank is information, not embarrassment.

Where an answer is genuinely unknowable — the regime is ambiguous, the crowding data is stale — writing “unknown” is the discipline working, and it argues for smaller size or no position at all. What the blank never licenses is skipping the question.

Why ten

Five questions miss the seam where most damage happens: the mismatch between the investor and the instrument. Horizon, loss tolerance, and reserve are exactly the questions a product pitch never asks, because they are about the buyer, not the product. Fifty questions fail in the opposite direction — a checklist nobody completes protects nobody.

Ten is the smallest set that covers the four surfaces where capital actually gets destroyed:

• the investor (questions 1–2)
• the environment (3 and 7)
• the portfolio’s structure (4–6)
• the position’s arithmetic and exit (8–10)

Remove any one and a documented failure class walks through the gap.

Where this breaks

Questions produce discipline, not foresight. Ten honest answers cannot make a bad year good; they make it survivable, and they eliminate the unforced errors — oversized positions, breached reserves, exits improvised in a falling market.

They also verify nothing: whether a claimed record is real belongs to the four-step claim audit, and whether the person presenting it deserves a hearing belongs to the advisor red-flag checklist. Answers decay, too — regime and crowding change, which is why every answer carries a date.

The working order for the three safeguards is by cost: run the red-flag checklist on the person, the claim audit on the number, and these ten questions on the decision. Capital moves only after all three — and the tenth answer is still allowed to be no.