Actionable insights — Quality always wins

1. Report the cash your portfolio earns, at a frequency short enough to feel

1. Compute your book's look-through free cash flow: your ownership share of each holding's free cash flow, summed.
2. Express it annually, then divide down — monthly, weekly, daily, hourly, per minute. The small denominations are the point: they make the accrual visible while the price does nothing.
3. Chart the figure over time and treat that line as the performance report you look at, with the market value as a secondary number.
4. State the reasoning explicitly so the metric cannot be used selectively: prices follow intrinsic value eventually, so a rising cash line is evidence the process is working even while the price disagrees.
5. Publish it in bad periods as well as good ones — a metric only introduced during a drawdown is a rationalisation, not a discipline.

• Look-through cash flow that grows only because you keep adding money — separate contributions from organic growth.
• The metric becoming a way to avoid a genuine thesis break; falling cash flow at a holding must still trigger a review.

2. Find the precedent with the numbers attached — including the underperformance figure

1. Identify a historical episode in which your style lost badly, and quantify it: index return, your style's return, and the gap in percentage points.
2. Choose managers whose process was later vindicated, so the example tests the strategy rather than the person.
3. Record both halves — the drawdown and the subsequent recovery — with dates, so the elapsed time is visible. Two years of pain is the datum, not a footnote.
4. Keep contemporaneous commentary (headlines, quotes) alongside the numbers; the social pressure is the part that is hardest to remember afterwards.
5. Frame the analogy honestly — "history doesn't repeat itself, but it often rhymes" — rather than claiming the mechanism must recur.

• Cherry-picked analogues; for every 1999 there is an episode where the losing style was simply obsolete.
• Using the precedent to defer portfolio-level work — the analogy is comfort, not evidence about your particular holdings.

3. Publish base rates before you need them

1. Write down, in advance and in plain numbers, the frequency of the bad outcomes your strategy will produce.
2. Cover three levels: the market (how often it falls 10%+), the individual position (how often a pick disappoints), and the relative result (how often you lag the index).
3. Communicate them when things are calm, so that when one occurs it is a forecast being confirmed rather than a surprise being explained.
4. Check your actual experience against the stated rates periodically. If disappointments are running at one in two rather than one in three, the process, not the market, is the problem.
5. Direct the message at the behaviour, not the portfolio — the failure mode is the holder capitulating, not the holding underperforming.

• Base rates broad enough to excuse anything; "one in three" is falsifiable, "sometimes" is not.
• A run of results well outside the stated rates being explained rather than investigated.

4. Treat a style switch at the extreme as the most expensive trade available

1. Notice when the switching impulse arrives — it correlates with the widest performance gap, not with new information.
2. Acknowledge the case for switching in full, without strawmanning it, so the decision is made against the strongest version.
3. Ask what would have to be true for the switch to work: the outperforming style must continue outperforming from an already extreme starting point.
4. Recognise the double cost — realising the loss on the abandoned style and buying the new one near its peak.
5. Pre-commit instead to a rule about when you would legitimately change approach (a broken premise, not a period of underperformance), and write it down now.

• The distinction between conviction and stubbornness — the rule you wrote down has to be capable of firing.
• Partial switching (a "small allocation" to the hot style) that quietly becomes the whole book.

5. Use the safest-versus-riskiest spread as a positioning gauge

1. Track the highest-risk cohort of the market and the lowest-risk cohort, each relative to the index.
2. Note when both hit extremes simultaneously in opposite directions — that is a positioning reading, not a valuation one.
3. Interpret it as a statement about crowding: an all-time low for safety means the marginal buyer has abandoned it entirely.
4. Do not time on it. Convert it into patience and, where the fundamentals justify, into buying — the same conclusion the earlier issues reached.

• Extremes that persist for years — "all-time low" is not a timing signal and can go lower.
• Definitions of "safest" that have drifted; check what is actually in the cohort.

6. Own the underperformance in writing before defending the process

1. State the bad result first, in the bluntest available terms, including your own responsibility for it.
2. Only then set out the evidence that the process is intact — competitive advantages, balance sheets, earnings, cash generation, item by item.
3. Be explicit about what you are not claiming: no timing, no prediction that the other side falls, no change in positioning.
4. End with the action, which in a drawdown letter is usually "no action" — but say so, so that inaction is a decision rather than a drift.

• The candour becoming the argument — an honest admission does not make the underlying analysis right.
• "Nothing has changed" claims that have not been re-tested holding by holding since the drawdown began.