ADVANCED: COURSE 2 | LESSON 1
The bias catalogue: loss aversion, revenge trading, overconfidence
Learning objectives
Connect the classic behavioural-finance findings (prospect theory, loss aversion, the disposition effect, overconfidence) to the specific trading errors they produce
Recognise the live symptoms of each bias in your own execution — cut winners, widened stops, revenge sequences, oversized "sure things"
Apply structural countermeasures that work because they remove the decision from the biased moment, not because you promise to be stronger
Your brain was not built for this
The uncomfortable premise of this course: the biases below are not character flaws that discipline erases. They are well-replicated features of human decision-making under uncertainty, documented across decades of research, and they show up in professional dealers and Nobel laureates as reliably as in first-year retail traders. The professional response is not "be stronger" — it is to design a process in which the biased decision never gets made in real time. Knowing the catalogue is step one; A2.2–A2.4 build the machinery.
Prospect theory and the ~2× asymmetry
Kahneman and Tversky's prospect theory (1979) — the work behind Kahneman's Nobel — replaced the rational-agent model with how people actually choose. Three findings matter directly for trading.
Losses loom larger than gains. Empirically, the pain of a loss is weighted roughly 2 to 2.25 times the pleasure of an equal gain. A trader who is up $400 and then back to flat feels worse than one who was never up at all, even though the accounts are identical.
Reference dependence. We evaluate outcomes as gains or losses relative to a reference point — usually our entry price — not as absolute wealth. This is why an open position at −30 pips feels like a wound to be healed rather than what it actually is: a fresh decision about the next move from the current price. The market does not know your entry.
Risk attitudes flip across the reference point. In the gain domain, people are risk-averse: offered a sure $500 versus a 50% chance of $1,100, most take the sure thing despite its lower expected value. In the loss domain, they become risk-seeking: offered a sure −$500 versus a 50% chance of −$1,100, most gamble to avoid locking in the loss. Now translate to a trading screen: the winner (gain domain) gets snatched early; the loser (loss domain) gets held, averaged into, or has its stop moved. Prospect theory predicts the single most common retail execution pattern before the trader ever opens a platform.
The disposition effect: the pattern in the data
That prediction has been measured. Shefrin and Statman named it the disposition effect; Terrance Odean's study of 10,000 retail brokerage accounts found investors were about 50% more likely to sell a winning position than a losing one — and, damningly, the winners they sold went on to outperform the losers they kept. The behaviour is not just emotionally costly; it is directionally wrong.
For a leveraged CFD trader the effect is more expensive still, because it systematically manufactures a bad R-multiple distribution: many small wins (cut at +0.5R "to be safe") against occasional catastrophic losses (the −4R that started life as a −1R with a stop that got moved). You can have a 65% win rate and lose money comfortably this way — recall the expectancy arithmetic from P2.1. When A2.3 has you plot your R-multiple histogram, a fat left tail plus a clipped right tail is the disposition effect's fingerprint in your own data.
Countermeasures that respect the bias: decide exit rules before entry, when you are in neither domain; make stop-widening physically inconvenient or impossible (some platforms allow it; your rules must not); manage trades by rule ("stop to breakeven at +1R, trail thereafter"), not by feel; and evaluate open positions with the question "would I open this trade now at this price?" — which forcibly resets the reference point.
Revenge trading: risk-seeking in the loss domain, live
Revenge trading is prospect theory with adrenaline. After a loss — especially one that feels unjust, a stop-hunt wick or a news spike — the trader is deep in the loss domain, risk-seeking, and focused on a reference point ("get back to where I was today") that the market does not share. The signature sequence: a loss, an immediate re-entry in the same or opposite direction within minutes, at larger size, with a thinner or absent setup. Each subsequent loss deepens the domain and enlarges the next bet. Most catastrophic single-day account losses are not one bad trade; they are a revenge sequence of four to six.
The tell is time and size: real setups arrive on the market's schedule, not 90 seconds after your stop-out. Countermeasures: a hard daily loss limit (e.g. −2R or −3R, then flat, platform closed — decided in your business plan, A3.4); a mandatory cooling-off period after any stop-out before the next order; and a rule that position size may never increase on the same day as a loss. These work because they are structural — they bind the calm you to protect against the tilted you.
Overconfidence and its accomplices
Barber and Odean's famous study title says it plainly: "Trading is hazardous to your wealth." Across 66,000 households, the most active traders underperformed the least active by around 6 percentage points a year — and their follow-up found men traded 45% more than women and earned less for it. Overconfidence expresses itself in trading as excessive frequency, oversized "conviction" positions, and abandoning tested rules after a hot streak ("I've evolved past the system").
Its accomplices complete the catalogue. Confirmation bias: once positioned, you rate supporting evidence up and disconfirming evidence down — the reason A1.5 makes you write invalidation conditions before the trade. Recency bias: the last five trades feel like the truth about your strategy; fifty is the minimum sample worth a feeling (A2.3 formalises this). Outcome bias: judging a decision by its result — the lucky unplanned winner from A1.5's quiz — which trains precisely the wrong habits. The hot-hand and gambler's fallacies, mirror-image errors about streaks in independent events: neither "I'm on fire, size up" nor "four losses, a win is due" has any statistical basis if trade outcomes are independent.
A sober note: awareness alone barely helps. Studies of debiasing consistently find that knowing about a bias does not reliably prevent it — the same machinery that generates the bias evaluates whether you are currently biased. That is why every countermeasure in this lesson is structural (rules, limits, pre-commitment) rather than motivational. The routine you built in A1.5 and the review system coming in A2.2–A2.3 are not administration around trading; they are the debiasing technology itself.
Key takeaways
Prospect theory: outcomes are judged from a reference point, losses weigh ~2× gains, and risk appetite flips — risk-averse in gains, risk-seeking in losses
The disposition effect (sell winners, ride losers) is measured fact — Odean found retail investors ~50% likelier to sell winners — and it manufactures a fat-left-tail R distribution
Revenge trading is loss-domain risk-seeking in real time; its antidotes are hard daily loss limits, cooling-off periods, and a no-size-up-after-loss rule
Overconfidence shows up as overtrading and rule abandonment; Barber & Odean showed the most active retail traders underperform materially
Awareness does not debias — structure does: pre-committed exits, written invalidations, and limits set by the calm version of you