Psychology & Process — Protecting the Plan From Yourself

Lesson 6 named the behavioral traps in a table. This lesson goes deeper: why process has to be judged separately from outcome, how each bias actually shows up in a brokerage account, and — most importantly — the mechanical antidotes that work when willpower does not: a pre-mortem, two checklists, a journal template, pre-committed sell rules, and a plan written for the day the portfolio is down 30%.

What you’ll learn

Why a good decision and a good outcome are different things — and the 2×2 matrix that keeps them apart
The eight biases that cost investors most, each with the specific mechanism that disarms it
The pre-mortem: how to surface the most likely failure before it can hurt you
Two checklists — one before a trade, one when you suspect the trade is emotional
An investment-journal template you can fill in in two minutes, and why closed positions get a post-mortem
The sell discipline as pre-committed rules, and the “reasons” that are not rules
How to size a drawdown in dollars before it happens, and how an information diet protects long-term positions

Process beats outcome

Markets are noisy. A careful decision can lose money and a reckless one can double — for a while. If you grade yourself only on outcomes you will learn the wrong lessons, because the feedback is mostly luck in the short run and only mostly skill over many decisions.

The fix is to grade two things separately: the quality of the decision at the time you made it (did you follow a sound process on the information available?) and the outcome (did it make money?).

	Good outcome	Bad outcome
Good decision	Deserved success. Repeat the process. Do not increase size just because it worked	Bad luck. The most valuable quadrant — resist the urge to “fix” a process that was right. Check whether the thesis actually broke or the price just moved
Bad decision	Dumb luck. The most dangerous quadrant — it teaches you that skipping the process is fine. Write down what you skipped before the money teaches you otherwise	Deserved loss. Painful but useful: the process failure is visible. Fix the process, not the stock

Everything below exists to move your decisions into the top row. Whether they land in the left or right column is only partly up to you.

The biases that cost the most — and their antidotes

Lesson 6 gave a one-line antidote per bias. Here is how each one actually appears in an account, and the mechanism — not the intention — that stops it.

Disposition effect

What it is: selling winners too early (to “lock in” the gain) and holding losers too long (to avoid “making the loss real”). It is loss aversion in action, and it is the single most documented pattern in retail trading records.

How it shows up: a portfolio where every remaining position is underwater and every closed one was a small gain.

Antidote (mechanism): the decision to sell is made by the thesis file, not by the P&L. If the thesis is INTACT, a gain is not a reason to sell; if the thesis is BROKEN, a loss is not a reason to hold. Write the exit triggers down at purchase; then the trade later is execution, not a decision.

Anchoring on the purchase price

What it is: treating what you paid as the reference point for every later decision — “I’ll sell when it gets back to even.”

Antidote: the one question that dissolves the anchor: “If I had this money in cash today, would I buy this stock at this price?” If yes, hold. If no, the purchase price is irrelevant — the market does not know what you paid.

Loss aversion

What it is: a loss hurts roughly twice as much as an equal gain feels good, so you act to avoid recognizing losses rather than to avoid incurring them.

Antidote: pre-commit to the exit rules (below) and size positions so that the worst case is survivable. A loss you sized for in advance is a cost of doing business; a loss you did not size for is a crisis.

Confirmation bias

What it is: reading only the bull case for what you own, and finding the bear case “already priced in.”

Antidote: run bear-case on your own holding, on purpose, before every scheduled review — and write down which of its Thesis-Killers you could not refute. If the answer is “none of them,” you have not read it honestly.

Recency and extrapolation

What it is: assuming the last three years are the next three; that a stock that has compounded at 30% “always does”; that the drawdown you are in will never end.

Antidote: zoom out to full cycles, and put a base rate next to every forecast: how many companies have sustained that growth rate for a decade? (Very few.) stock-valuation’s reverse-DCF asks what the price requires — recency bias is usually visible there.

FOMO and herding

What it is: buying because it is going up and everyone is talking about it — after most of the move.

Antidote: a rule that a position can only be opened from a thesis written before looking at the price chart or the social feed. If the thesis was written after the excitement, it is a rationalization. Cooling-off rule: 24 hours between the idea and the order.

Overconfidence

What it is: rating your own accuracy higher than your record supports, and sizing accordingly.

Antidote: position caps that do not move with conviction — a hard maximum per name regardless of how sure you feel — and a journal that shows your actual hit rate.

Sunk cost — averaging down on a broken story

What it is: adding to a losing position because it has lost, to “bring the average down,” when the reason you bought it no longer holds.

Antidote: two mechanisms working together — a hard share-count ceiling set at purchase (position-ladder), and a thesis gate: no adding while the thesis is WEAKENED, none at all when BROKEN (thesis-tracker). Averaging down inside an intact thesis and a stated ceiling is a plan; outside them it is the disposition effect with leverage.

The pre-mortem

A post-mortem explains a loss after it happened. A pre-mortem imagines it first, while you can still wire the cause into a trigger.

Write one paragraph, in the past tense, dated twelve months out:

“It is [date + 12 months]. This position lost 40%. Looking back, the reason was …”

Then read what you wrote and ask: is there a trigger in my plan that would have caught that before −40%? If the pre-mortem names a cause no trigger covers, add the trigger. Common findings: “the customer concentration I dismissed”, “the multiple compressed before the earnings did”, “I kept adding on the way down.”

The pre-mortem works because it converts vague optimism into a specific failure path — and because it is far easier to imagine a loss when you have not yet lost anything.

Two checklists

Before a trade

[ ] The thesis is written in one paragraph: what has to happen, why the market is wrong, why now
[ ] 3–5 KPIs with thresholds, and the triggers that would make me exit
[ ] I have read the bear case and can say which argument worries me most
[ ] Position size is inside my cap, and I know the dollar loss at −30% and −50%
[ ] The pre-mortem is written
[ ] The order is a limit order, and the price is one I would still pay tomorrow
[ ] The journal entry is drafted before the order, not after

“Am I about to make an emotional trade?”

Stop and answer honestly if any of these is true:

[ ] The stock moved more than 10% today and that is why I am here
[ ] I have checked the price more than three times today
[ ] I am acting on a headline I read in the last hour
[ ] I want to “get back to even” or “not miss it”
[ ] I cannot state which trigger in my plan fired
[ ] I am about to add to a position whose thesis is WEAKENED or BROKEN
[ ] I feel relief or excitement about the trade rather than nothing in particular

Two or more ticks: wait 24 hours, then re-run the before a trade checklist.

The investment journal

The journal is where the process-vs-outcome matrix becomes data. Two minutes per entry; one entry per action (open, add, trim, close, and every scheduled review).

Date            | 2026-05-30
Ticker · action | NVDA · open
Price · size    | $118 · 40 sh (3.8% of portfolio; cap 5%)
Thesis (1 line) | Data-center demand compounds for 2+ years; market prices a one-off cycle
Proves me wrong | Gross margin < 70% · DC growth < 40% YoY · top-3 customer announces in-house silicon at scale
Emotional state | 2 / 5  (1 = calm, 5 = urgent)
Informed by     | stock-eval, bear-case, stock-valuation — 2026-05-28 run
Next check      | Aug print

Rules:

Emotional state is not optional. Over a year, the entries scored 4–5 will show you exactly which trades to stop making.
Every closed position gets a post-mortem using the matrix: which quadrant, what the process did well or skipped, one sentence of lesson. Keep the closed files — they are the only honest record of your hit rate and your average winner vs. average loser.
Review the journal, not the portfolio, at each scheduled check.

The sell discipline: rules vs. reasons

Selling is where most damage happens, because it is where emotion has the most leverage. Pre-commit to the rules and refuse the reasons.

Rule (sell, or seriously consider it)	Not a rule (feelings dressed as reasons)
The thesis is broken — a KPI crossed its threshold or an exit trigger fired	“It’s down 20%”
The target is reached and it is no longer cheap — the valuation range from `stock-valuation` is exceeded with fundamentals unchanged	“It’s up 40%, I should take profits”
A better use of the capital — a clearly superior opportunity that you would fund from this position specifically	“I’m bored with it”
The position is too large — it has grown past the cap and the concentration risk is now the story	“There was a scary headline”
You need the money — the goal the money was for has arrived	“Everyone is selling”

The left column can all be checked against a document you wrote when calm. The right column cannot.

Drawdowns: size the pain before it arrives

Every long-term equity portfolio will fall 30% at some point; many will fall 50%. The question is not whether, but whether you have already decided what you will do.

Translate the percentage into dollars, today. A $200,000 portfolio at −30% is $140,000 — a $60,000 loss on the screen. Write that number down. If seeing it makes you want to change the allocation, change it now, in calm, rather than on the day.

The plan is written for that day. The pre-trade checklist, the exit rules, and the thesis files are not for ordinary Tuesdays; they exist so that on the day the account is down $60,000 you execute a decision you already made rather than make a new one under stress.

A drawdown is not a trigger. A 30% fall in price with every KPI inside threshold is a price event, not a thesis event. The thesis file tells you which one you are looking at. risk-stress-test lets you see the 2008 / 2020 / 2022 numbers on your own holdings before you own them.

Information diet

More information does not produce better decisions past a low threshold; it produces more trades. For long-term positions:

Scheduled reviews, not continuous monitoring — the thesis file’s next-check date, earnings, and quarter-end. Nothing in between unless a trigger fires.
No intraday price checks for positions with a multi-year horizon. The price today carries no information about a thesis measured in quarters.
Fewer, better sources — the filings, the call, one or two analysts you have tested against outcomes. Not the feed.
Read the bear case at least as often as the bull case. If your inputs are all one direction, that is the diet talking.

A tiny worked illustration (illustrative, rounded)

Two positions, both bought at $100, both held one year.

	Stock A	Stock B
Price now	$140 (+40%)	$70 (−30%)
Thesis status	INTACT — every KPI improving; fair value range $150–180	BROKEN — the core customer left; two KPIs breached
Disposition-effect action	Sell A (“lock in the gain”) · Hold B (“it’ll come back”)
Process action	Hold A (still cheap vs. fair value, thesis intact) · Sell B (thesis broken; the purchase price is not a reason)

The disposition-effect portfolio keeps the broken company and gives away the compounder. The process portfolio does the opposite — and never had to make a decision on the day, because both rules were written at purchase.

How the plugin helps

In the plugin: thesis-tracker is the written contract — it stores the thesis, KPIs, triggers, pre-mortem, and decision log, and returns INTACT / WEAKENED / BROKEN so the sell discipline is a lookup, not a feeling. bear-case red-teams your own holding on schedule. position-ladder sets the share-count ceiling that makes averaging down bounded. risk-stress-test puts the drawdown in dollars before it happens. learning-coach explains any output and asks “what would change your mind?”; result-validator checks that the analysis you are acting on is sound.

Check yourself

You followed your full process on a stock and it fell 25% because the whole sector sold off; every KPI is inside threshold. Which quadrant of the matrix is this, and what should you not do?
A friend says “I’ll sell when it gets back to what I paid.” Which bias is this, and what single question dissolves it?
Name the two mechanisms that together make averaging down a plan rather than a mistake.
Write a one-sentence pre-mortem for a stock whose top customer is 30% of revenue.
Of these five, which are sell rules and which are reasons: “the exit trigger fired”, “it’s up 40%”, “the position is now 12% of the portfolio against a 5% cap”, “a scary headline”, “everyone is selling”?

Answers

Good decision / bad outcome. Do not “fix” the process or sell because of the price alone — check whether the thesis broke (it did not) and hold to the plan.
Anchoring on the purchase price. “If I had this money in cash today, would I buy this stock at this price?”
A hard share-count ceiling set at purchase (position-ladder) and a thesis gate — no adding while WEAKENED, none when BROKEN (thesis-tracker).
Example: “It is 12 months later and this lost 40% because the top customer, 30% of revenue, moved the contract in-house and the multiple compressed before the next quarter reported.” → add a customer-concentration trigger.
Rules: the exit trigger fired; the position is 12% against a 5% cap. Reasons (not rules): it’s up 40%; a scary headline; everyone is selling.

Key takeaways

Grade decisions and outcomes separately; the 2×2 matrix keeps luck from teaching you the wrong lesson.
Each bias has a mechanism that beats it: the thesis file beats the disposition effect, “would I buy it today?” beats anchoring, bear-case beats confirmation, a hard ceiling plus a thesis gate beats averaging down on a broken story.
Write the pre-mortem before the order; if it names a failure no trigger catches, add the trigger.
Use the two checklists — before a trade, and whenever a trade feels urgent. Two ticks on the emotional checklist means wait 24 hours.
Keep a journal with an emotional-state score, and post-mortem every closed position.
Selling follows pre-committed rules (thesis broken, target reached and no longer cheap, better use of capital, position too large); “it’s down”, “it’s up”, “headline”, “boredom” are not rules.
Size the drawdown in dollars before it arrives; the plan exists for that day. Review on a schedule, not on the price.

Next / Related: Previous lesson — Earnings Season, Explained. Next: Case Study: When the Answer Is No shows the loop ending in a pass, twice. Or head back to the Learning hub, then Choose a Skill to put it to work. See also Concepts and the Glossary.

Educational content only. Not financial advice.