It takes a Sunday afternoon and most of an evening. You read the annual report, you read the two before it, you work out roughly what the business earns and what you are being asked to pay for it, and at about eleven o’clock you write a line in the notes app: buy this, ₹20,000. You do not act, because the salary arrives on the first and the money is not there. On the first there is a school admission fee nobody had budgeted for. On the fifteenth of the following month you remember, open the app, see that the price is 19 per cent higher, and buy anyway — because the decision was already made. That last clause is the whole lesson. Nothing was reread. The conclusion made it through six weeks intact, because a conclusion is one sentence and one sentence is easy to carry; the three pages of reasoning that produced it did not, because nobody carries three pages. What arrived at the order screen on the fifteenth was not a decision. It was an instruction, and instructions do not argue back.
A doctor writes a prescription for a particular fever in a particular week, after asking questions and looking at you. Seven months later somebody in the house has a fever, the paper is still in the drawer, and the strip gets taken out — because a doctor prescribed it. Nobody thinks the prescription is a lie. It is a perfectly accurate record of a conclusion reached about a different week.
Your note said buy, at ₹430, for three reasons you could have recited in April. By the middle of May it says buy. The price is gone, the reasons are gone, and what is left is the one part that was small enough to remember — which is also the only part that cannot be checked against anything.
What a decision is made of, and which parts survive storage
| The part | What it was on the day | How long it survives in your head |
|---|---|---|
| The conclusion | Four words: buy this, ₹20,000 | Indefinitely. It is short, it is portable, and it feels like the decision itself |
| The price it was true at | ₹430 | Days. And once it is gone, today’s price is accepted as though it were the one you tested |
| The reasons | Three pages of notes and a spreadsheet | A fortnight, generously. What actually remains is a feeling of having done the work |
| What would make it wrong | Almost never written down at all | Nothing to survive. This is the part that would have turned the note back into a decision when you reread it |
| The size | ₹20,000, which was 4% of the portfolio in April | Survives as a rupee figure and quietly stops meaning what it meant, because the portfolio it was 4% of has changed |
The three ways a note goes stale
- The price moved. The easiest of the three to check, and the one people most reliably decline to check, for a reason that is worth saying out loud: checking might cost them the decision, and by the fifteenth they want the decision. That is [[confirmation bias]] doing its ordinary work — not manufacturing evidence, simply declining to look for the one number that could end the matter.
- A fact arrived. Quarterly results, a change of chief executive, a regulatory order, a promoter share pledge, an auditor resigning. Exchange filings are continuous and your note is not. Nothing in a notes app updates itself when a company files something.
- You moved. The 4% position sized in April against a portfolio of ₹5,00,000 is a 4% position no longer if a bonus arrived, an income stopped, or you added ₹3,00,000 elsewhere. The company can be entirely unchanged and the decision still be wrong-sized, because [[position sizing]] is a statement about you rather than about the stock.
Why the delay is not the problem
The instinct after reading this far is to resolve to act faster, and that is the expensive misreading. The gap between deciding and acting is frequently the most useful part of the process — it is where a bad idea dies of boredom, where the enthusiasm of an evening fails to survive a Wednesday, where a tip stops sounding urgent. An earlier module in this track spends its length on how many of the clocks you feel are manufactured by somebody who profits from your haste. Nothing here contradicts that. The problem is not the interval. It is that nothing happens during it. A note left alone for six weeks is either re-derived or obeyed, and the difference between those two outcomes is one short paragraph written on the night you make the decision.
Writing a decision so that it can age
- 1The price this was true at, and how far it can move
One number, and a band around it — 8%, 15%, whatever suits the case. The figure matters far less than the existence of a figure, because a written band forces the comparison that memory refuses to make. This is [[margin of safety]] expressed as a shelf life rather than as a discount.
- 2Two or three things that would make me drop this
Observable ones, checkable in a filing: margins below a level, debt above one, the division being sold. This is the line that converts a stored conclusion back into a decision when you reread it, because it gives the reread something to do.
- 3The size, as a share of the portfolio
Write 4%, not ₹20,000. A percentage re-sizes itself against whatever the portfolio has become; a rupee figure is remembered perfectly and silently means something else.
- 4Review by — a date
After that date the note is not a decision at all. It is a research file, which is a genuinely valuable thing to own and a completely different thing to act on. Undated notes never die; they simply get executed at an arbitrary later moment by a version of you with less information and more impatience.
The note you find in October
In the notes app is a line from July: "Buy ₹35,000 of this. Cheap on cash flow, and the debt is coming down." The price today is 6% above where it was in July. You have ₹35,000 spare and twenty minutes.
A note written in March said a stock was worth buying at ₹610, because it traded at roughly 14 times earnings against a sector on about 20. In August, with earnings unchanged, the stock is ₹790 and the note is acted on as written. What has actually happened?
April ki raat notes mein likha: "yeh khareedna hai, ₹20,000." Paisa nahi tha, phir school ki fees aa gayi, aur agle mahine ki 15 ko aakhir khareed liya — bina dobara padhe, kyunki "faisla toh ho hi chuka tha". Ek line yaad rah gayi, teen panne bhool gaye. Bhaav ₹430 se ₹512 ho gaya, yaani 19% upar, aur kamai wahi ki wahi — toh jo 18 guna tha woh ab 21.4 guna hai. Poora case 18 aur 24 ke beech ke chhe point ka tha; ab sirf 2.6 point bacha hai. Ho sakta hai itna bhi kaafi ho — par kisi ne poocha hi nahi. Jo faisla aaj ke numbers se dobara nikala na ja sake, woh faisla nahi, order hai.
- A stored decision decays into an instruction — the conclusion survives being remembered and the reasons do not.
- Rising prices consume the case for buying; they do not confirm it.
- Write the price it was true at, what would make it wrong, the size as a percentage, and a review date.
- Size written in rupees stops meaning what it meant; size written as a share of the portfolio re-sizes itself.
- The delay is not the defect. The defect is that nothing is asked to happen during it.
Mark it done to track your progress through the curriculum.