Side Lesson 15: Trading Psychology — The Seven Biases That Bleed Money

---

1. Why This Is Important

You can run every spreadsheet in this course, you can quote the Sharpe ratio of every factor, you can recite Damodaran's equity-risk premium update from memory — and you can still lose money. The reason is that the keyboard is connected to a biological brain that evolved to keep a primate alive on a savannah, not to underwrite cash-flow risk over thirty years. Trading psychology is the discipline of spotting your own brain's defaults and overriding them with rules. Four reasons to take it seriously:

The behaviour gap is real, even if the exact number is debated.

Dalbar's Quantitative Analysis of Investor Behavior has reported for thirty years that the average equity-fund investor earns 3-5 percentage points per year less than the funds they own. The methodology has been picked apart (Vanguard, Morningstar, and academics have written critiques) — and even the corrected, conservative reads still find a 1.5-2 pp/yr drag from poorly-timed buys and sells. Compounded across a 30-year career on a $100k starting balance, a 3 pp/yr gap turns roughly $1.0M into roughly $400k. The gap pays for someone else's vacation home.

Bias is invisible in real time. Confirmation bias does not

announce itself. You do not feel anchored to your buy price; you feel rational. The whole problem of behavioural finance is that the brain's bugs feel like the brain's normal output. The only defence is structural — pre-commitment, written plans, automation — because in-the-moment introspection is exactly when you cannot trust yourself.

The damage compounds asymmetrically. Per Kahneman-Tversky

prospect theory, a $100 loss feels roughly twice as bad as a $100 gain feels good (lambda ~ 2.0-2.25). That asymmetry biases every decision: you cut winners early to lock in the good feeling, you hold losers because realising them hurts. You end up with a portfolio of losers — the disposition effect, the most documented bias in retail brokerage data.

The market is a beauty contest. The market can

stay irrational longer than you can stay solvent. Every bias that blows up an individual blows up more spectacularly when you have leverage, a margin call, or a deadline. Behavioural discipline is not a soft skill; it is the precondition for compounding.

Side 15 lays out the seven biases that do the most damage, gives the counter-measures that actually work, and shows the cumulative wealth gap from buying-and-selling-emotionally on real S&P data.

The interactive lab on the website lets you walk through three scenario decisions — stock down 30% — sell or hold?, *stock up 50% — sell or hold?, missed the rally — chase or wait?* — and read both the rational answer and the typical behavioural answer side by side.

---

2. What You Need to Know

2.1 Loss Aversion — The 2:1 Asymmetry

Daniel Kahneman and Amos Tversky's 1979 prospect theory paper, and the 1992 cumulative refinement, established the empirical *value function*:

$$ v(x) = \begin{cases} x^{0.88} & x \geq 0 \\ -\lambda \, (-x)^{0.88} & x < 0 \end{cases} $$

The two pieces of that function tell you almost everything about investor behaviour:

• Concave in gains. The pleasure of going from $1,000 to $2,000

is bigger than the pleasure of going from $10,000 to $11,000, even though the dollar gain is identical. Diminishing marginal utility.

• Convex in losses, AND steeper. The pain of losing $100 is

roughly twice the pleasure of gaining $100 (lambda ~ 2.0-2.25 in most replications). The kink at zero is the entire reason markets have crashes — no one wants to realise a loss, until everyone decides to realise the loss at once.

Loss aversion in practice produces three predictable mistakes: holding losers too long (because realising the loss hurts), selling winners too early (the disposition effect), and being too conservative about position sizing right after a drawdown. The fix is not to feel less pain — you can't — but to pre-commit to the sell rule before the loss happens.

2.2 Recency Bias and Performance-Chasing

Recency bias is the brain's habit of weighting the last 12 months as if it were the eternal future. After a year like 2023 (S&P +26%) or 2009 (S&P +27%), retail flows pile into US large-cap. After 2008 (-37%) or 2022 (-18%), retail flows leave. Morningstar's flow data and the Investment Company Institute time-series both show the same pattern: buy-high, sell-low at scale.

The math punishes recency. A simple back-of-envelope: if you switch from cash to equities after a +20% year and switch back to cash after a -20% year, you systematically miss the mean-reversion that follows both extremes. Damodaran's 1928-2024 dataset shows that the five years following a -20% calendar year averaged +14%/yr and the five years following a +30% year averaged +6%/yr. The recency-driven investor catches the wrong tail of every distribution.

2.3 Confirmation Bias and Overconfidence

Confirmation bias is the tendency to seek and over-weight information that supports your existing position, while ignoring or rationalising contrary evidence. You bought TSLA at $400, the price falls to $200, and you find five articles explaining why this is a buying opportunity and zero articles explaining why your thesis is broken. The brain's filter is doing its job — keeping the existing model stable — and the portfolio pays for it.

Overconfidence is the close cousin and gets worse after streaks. A 2001 study by Brad Barber and Terrance Odean (Boys Will Be Boys) showed that men trade 45% more than women and underperform by ~1.4 pp/yr; the gap is mostly explained by overconfidence after winning trades. After a few wins, the brain attributes the wins to skill and the losses to luck — even when the actual ratio is reversed. The fix: pre-trade journalling (write the thesis, the entry, the stop, and the time horizon before pressing the button) so that post-hoc rationalisation has a paper trail to argue against.

2.4 Anchoring, the Disposition Effect, and FOMO

Three closely-related biases that warp the exit decision:

• Anchoring to the buy price. The buy price has zero economic

meaning to anyone except your tax form. The market does not know what you paid. But your brain treats "I bought at $80" as a magic reference point — green above $80, red below $80. The right reference is forward expected return from here, not "did I make my money back yet." If you were handed the position today as a gift with no cost basis, would you hold it? If no, sell.

• Disposition effect. The empirical pattern, documented by

Hersh Shefrin and Meir Statman in 1985 and replicated in every retail brokerage dataset since, is that investors realise gains about 1.5x as often as they realise equally-sized losses. The result is a portfolio that is, on average, a collection of stocks that have gone down since purchase — the worst possible selection mechanism.

• FOMO and revenge trading. FOMO (fear of missing out) is

recency bias dressed up as opportunity: the market ran 30% in three months without you, you chase, the rally tops, you take the drawdown. Revenge trading is the doubly-broken cousin — taking a bigger position right after a loss to "get it back," which converts a manageable drawdown into a catastrophic one. Said another way: you do not get to negotiate with the market.

2.5 Anchoring's Cousin: The Gambler's Fallacy

The gambler's fallacy is the belief that independent outcomes are due to balance — five red roulette spins in a row "must" be followed by black. In markets, it shows up as: "the S&P has gone up five years in a row, it's due for a crash." The market does not owe you a crash, and prior calendar-year returns have near-zero predictive power on the next calendar year's return (the auto-correlation at the annual horizon is statistically indistinguishable from zero in the 1928-2024 series). Long horizons do mean-revert (10-year forward returns are negatively correlated with 10-year trailing returns at r ~ -0.4) but one-year horizons do not. Sizing trades from "we are due" is a coin-flip dressed up as analysis.

2.6 Behavioural Counter-Measures That Actually Work

The literature is clear: in-the-moment willpower does not work, but structure does. Six counter-measures, ordered by leverage:

Pre-commitment, in writing. Before the trade, write down

the thesis, the entry price, the stop level, the size, and the review date. Sign and date it. The behavioural literature (Ariely, Thaler, Sunstein) is unanimous: commitments made when you are calm bind you when you are emotional. The useful format: one paragraph for the thesis, one number for the size, one number for the stop.

Automation. Auto-deduct contributions to your IRA and 401(k)

every paycheck. Auto-rebalance the index fund sleeve on a date schedule (quarterly or annually) or a drift-band trigger (rebalance when any sleeve drifts more than 5 pp from target). Automation removes the moment of decision, and the moment of decision is when you make the mistake.

Checklists. A 6-item pre-trade checklist (thesis written?

size below 5% of portfolio? correlation to existing positions below 0.7? options assignment understood? tax lot identified? stop-loss set?) catches 80% of impulse trades. Borrow the format from Atul Gawande or the airline industry — the evidence base is the same as for surgery.

Kill switches. Set a max daily loss (e.g., 2% of account) and

a max weekly loss (e.g., 5%). When hit, you flat the active sleeve and walk away from the screen for 24 hours. The kill switch works precisely because it is binary and pre-committed — no judgement call required when you are already tilted.

The 24-hour rule. For any new position above 2% of the

portfolio, write the thesis today and place the trade tomorrow. 24 hours is enough for most FOMO impulses to dissipate. If the thesis still holds tomorrow, the trade is probably real. If it doesn't, you saved yourself a bad print.

Position-sizing rules. No single position above 5% at cost,

no sector above 25%, no leverage above 2.0x portfolio-wide. These are barbell anchors that keep one bad print from ending the run. From the Kelly literature: even if your edge is real, half-Kelly is empirically optimal because the brain's probability estimates are biased upward.

2.7 The Compound Cost of the Behaviour Gap

Take the conservative end of the academic range: a 2 pp/yr drag from behaviour on a $100,000 starting balance, 30 years, vs an 8%/yr buy-and-hold benchmark. The math:

$$ \text{B\\&H: } \$100{,}000 \cdot 1.08^{30} = \$1{,}006{,}266 $$

$$ \text{Behavioural: } \$100{,}000 \cdot 1.06^{30} = \$574{,}349 $$

The gap is $432,000 — more than four times the starting balance, in forfeited terminal wealth, from a drag that *feels invisible year-to-year*. At the Dalbar headline number of 3-5 pp/yr the gap becomes catastrophic ($1.0M vs ~$330k at 4 pp/yr). The behaviour gap is the single largest fixable cost in retail investing — larger than expense ratios, larger than tax drag (Side 4), larger than fund selection. Fix it first, then optimise everything else.

---

3. Common Misconceptions

"I'm too rational for this to apply to me." The literature

says no. Lab experiments find that more sophisticated investors are if anything more susceptible to overconfidence, because their priors are more strongly held. The fix is structural, not intellectual.

"The Dalbar number is fake — I read the methodology critique."

The headline 3-5 pp/yr number is fragile and over-states the gap for the median investor. But every honest replication (Vanguard's "Putting a value on your value", Morningstar's "Mind the Gap" annual study) finds a 1.0-1.7 pp/yr drag from poor timing. The number is real, just smaller than Dalbar prints. At 1 pp/yr over 30 years that is still a $200k loss on $100k starting.

"Loss aversion just means I'm conservative." No. Loss

aversion makes you more aggressive on losing positions (you refuse to sell, you double-down to recover) and more conservative on winning positions (you book the gain to lock in the good feeling). It distorts both sides asymmetrically.

"If I just knew more, the biases would go away." Education

reduces some biases (gambler's fallacy, base-rate neglect) but amplifies others (overconfidence, confirmation bias). Knowing you have a bias does not protect you from acting on it in real time. Pre-commitment does.

**"Dollar-cost averaging is just behavioural — the math says lump

sum wins."* Dollar-cost averaging is consciously* trading 0.5-0.7 pp/yr of expected return (per Vanguard's 2012 / 2023 studies) for guaranteed removal of regret risk. If the alternative is "stare at the lump sum, panic, and never deploy," DCA's expected-value loss is small relative to the expected-value loss of doing nothing. Behavioural insurance has a price.

"Stop-losses are for traders, not investors." Stop-losses on

individual stocks can be helpful if you size them at 2-3 ATR (well outside daily noise). Stop-losses on index funds are destructive because indexes mean-revert. The right tool for index exposure is position sizing, not stops — keep the index sleeve small enough that you can hold through a 50% drawdown without capitulating.

**"Revenge trading is just bad discipline — it's not a real

bias."** It is a real bias. Brain-imaging studies show that the loss-of-money signal activates the same threat regions as physical pain, and the limbic response biases toward immediate risk-seeking. The brain is engineered to chase losses; calling it bad discipline is like calling fight-or-flight bad discipline. The fix is the kill switch, not willpower.

"FOMO is fine when the rally is real." The problem with

FOMO is not that the rally is fake — sometimes it isn't. The problem is that you size the position by the strength of the feeling, not by the strength of the evidence. A real rally with a sized-too-large position becomes the next drawdown disaster.

**"Meditation / journaling / breathing exercises won't move the

needle."** Self-reported and brokerage-paired studies disagree. Investors who keep a written trade journal show measurably smaller behaviour gaps in the follow-up data. The mechanism is not the journaling itself — it is that journaling forces pre-trade pre-commitment.

**"The behaviour gap is just for active traders — index

investors are immune."** Wrong. The Dalbar gap is largest in active fund flows, but the same study shows a 1-2 pp/yr gap even among index-fund investors who panic-sell during drawdowns. The vehicle is not the protection; the behaviour is.

---

4. Q&A Section

**Q1. If loss aversion is hard-wired, what's the point of learning about it?** You can't turn it off, but you can route around it. Pre-commitment (write the stop before you enter), automation (rebalance on a schedule, not on a feeling), and position sizing (small enough that a loss isn't existential) all attack the consequences of loss aversion without requiring you to feel less pain. The brain stays broken; the portfolio is fine.

Q2. How big is the behaviour gap in dollars? Conservative (1.5 pp/yr) on $100k starting, 30y, 8% benchmark: ~$1.0M vs ~$650k → $350k forfeited. Aggressive (4 pp/yr per Dalbar): ~$1.0M vs ~$330k → $670k forfeited. Either way, larger than any other fixable cost in retail investing.

Q3. What's the single most useful behavioural counter-measure? Automation — specifically, paycheck-deducted contributions to an IRA / 401(k). It removes the decision moment entirely. Every other counter-measure (checklists, journals, kill switches) is incremental; automation is structural.

Q4. Should I check my portfolio less often? Yes. The behavioural finance literature broadly finds that more frequent portfolio checking → more loss-realisation pain (loss-aversion fires) → more bad trades. Quarterly is enough for most retail investors. Daily is actively harmful.

Q5. The Dalbar number is suspect — should I ignore the behaviour gap? No. The number is suspect at 4-5 pp/yr; it is robustly defensible at 1-2 pp/yr. Even at the conservative end the cost over 30 years is multiples of the starting balance. The fix is the same regardless of the exact magnitude.

**Q6. How do I handle a position that's down 30% and I think the thesis is still right?** Re-read the original written thesis (you wrote one — see §2.6). If the thesis is intact and the size is below 5% of the portfolio, hold. If the thesis broke (the company missed earnings, the catalyst evaporated, the risk you didn't anticipate showed up), sell — even at the 30% loss. The buy price is irrelevant; the forward expected return is everything.

**Q7. What's the difference between conviction and confirmation bias?** Conviction is a position you hold *while actively looking for disconfirming evidence* and updating your sizing on what you find. Confirmation bias is the position you hold *while filtering out disconfirming evidence*. The test: can you state, in writing, the specific data point that would make you sell? If yes, conviction. If no, bias.

Q8. Are kill switches realistic for retail investors? Yes — but easier with brokerages that let you pre-set them (IBKR, ThinkOrSwim). Manual kill switches require more discipline ("if my account is down 5% this week, I close all active positions on Monday morning"), but they are still better than nothing. Pre-committed in writing, reviewed weekly.

Q9. How does the irrational-longer-than-solvent rule relate to all this? That rule — *the market can stay irrational longer than you can stay solvent* — is the meta-rule. Every behavioural bias gets amplified by leverage and by deadlines. If your time horizon is 30 years and you have no leverage, you can survive your own brain. If you have 3:1 leverage and a margin clock, your brain's biases will compound into a forced liquidation before the thesis is right. Strict position sizing is the cheapest behavioural insurance you can buy.

Q10. What role does meditation / breathwork play? Marginal but non-zero. The mechanism, where it works, is reducing the amplitude of the limbic response to a loss — so the kill switch fires before the revenge trade. It is a complement to structural counter-measures, not a substitute. If you have to choose: write the thesis first, breathe second.

**Q11. Can I just hire someone to manage this for me to avoid the biases?** Partially. A fee-only fiduciary advisor at 0.5-1.0%/yr is buying you one counter-measure — the friction of having to call them before trading. Worth it for some people, not for others. Robo-advisors at 0.25%/yr buy you automation (the most useful counter-measure) without the human friction. Either is better than self-managing without a structure.

Q12. What's the one thing I should do tomorrow? Set up a paycheck-deducted IRA / 401(k) contribution at the highest amount you can afford, automatically rebalanced annually. Everything else in this lesson is incremental compared to that one structural change. Automate the contribution; the contribution is the trade you can never time wrong.