02 · A Field Guide to Cognitive Biases in Trading
Article 01 covered the theory: why people aren't rational. This one holds up the mirror — the 12 most common cognitive biases in trading, each with a definition, a trading example, and countermeasures. You don't lack willpower; you've been hit by these biases. And the scariest part of a bias is this: when you're making the mistake, you always feel certain you're right.
💀 Iron Law: When You're Making the Mistake, You Always Feel You're Right
When you're making the mistake, you always feel you're right. So the scariest thing isn't "making an error" but "not realizing you're making one" — while you believe you're analyzing rationally, anchoring, confirmation bias, and hindsight bias may all be running at once. Only checklists, rules, and review mechanisms can make invisible biases visible.
I. Decision-Making Biases
1.1 Anchoring (Anchoring)
- [What it is] Over-relying on the first piece of information received (the anchor) when judging; later information gets "adjusted" around the anchor, but never enough.
- [Trading example] You buy BTC at 60,000; it drops to 50,000 and you refuse to cut, because "I'll exit when it's back to 60,000" — you're in love with your own cost basis. Seeing "all-time high 100, current price 80" makes it feel cheap, so you ignore deteriorating fundamentals — 80 is merely cheaper than 100, not cheap in itself. Analyst target prices and other people's call prices are all anchors planted for you.
- [Countermeasure] Let stop-loss/take-profit levels be determined only by technical structure or risk budget — never by cost basis. Before entering, run a "flat-position test": "If I held no position right now, would I make the same decision at this price?" Hide average entry price and unrealized P&L from your interface.
1.2 Representativeness Heuristic (Representativeness Heuristic)
- [What it is] Substituting "does it look like" for "is it": extrapolating the character of recent price action into a long-term law.
- [Trading example] A coin that has risen for a month: "the trend is established, it'll keep rising." A stock down three days straight: "it'll surely keep falling." "Looks like a bull market" is not the same as "is a bull market." Using the last 30 days to represent the market's long-run distribution is using a sample to replace the population.
- [Countermeasure] Give "trend" an operational definition (e.g., price above moving averages + higher highs and higher lows together) and act only when the definition holds; look at statistics over long horizons, not just the recent stretch.
1.3 Availability Heuristic (Availability Heuristic)
- [What it is] Estimating probability by "how easily examples come to mind" rather than true frequency. Easily recalled events get overweighted.
- [Trading example] Hearing too many stories of friends getting rich at 10x leverage makes you overestimate get-rich odds and underestimate liquidation odds. Media covering a coin's surge daily makes it feel like it's "going mainstream." What memory retains best is precisely the rare extreme cases.
- [Countermeasure] Replace story-memory with real statistics (your P&L since opening the account, your strategy's historical win rate); remind yourself: behind every get-rich story you've heard lie a hundred liquidations you haven't.
1.4 Confirmation Bias (Confirmation Bias)
- [What it is] Seeking out and believing only information that supports existing views, while ignoring, discounting, or attacking contrary evidence.
- [Trading example] After going heavily long, every bearish item reads as "bad news fully priced in" or "market-maker shakeout." The same candle looks like a "dip-buying spot" to bulls and a "bounce-selling spot" to bears — conclusion first, evidence later; trading becomes self-persuasion.
- [Countermeasure] Force yourself to write at least 3 reasons against the trade before entering; if you can't, you haven't thought it through. In reviews, specifically tally "contrary signals I ignored." Base trend judgment on objective data — price, volume — and treat news opinions as background only.
1.5 Hindsight Bias (Hindsight Bias)
- [What it is] After events unfold, firmly believing "I knew it all along" — in hindsight everything looks inevitable.
- [Trading example] "I said we should have sold at the top," "should have known not to buy that dip." It creates two illusions: ① overestimating your judgment (bigger positions next time); ② underestimating luck in outcomes (a bad decision that made money = a good decision).
- [Countermeasure] Trading journals must record "judgment, reasoning, and emotion at decision time"; review against them afterward and forbid rewriting reasons based on results. Regularly ask: if the outcome had been opposite, how would I have explained it then?
II. Position & Exit Biases
2.1 Sunk Cost Fallacy (Sunk Cost Fallacy)
- [What it is] Continuing to invest because "so much is already in," ignoring the marginal expectation of further investment. Past costs shouldn't affect current decisions — but people can't help it.
- [Trading example] Down 40% and averaging down: "I'm already this deep — how else do I break even?" Holding a position whose trend is clearly broken because "cutting now hurts too much." The 40% you lost is already gone. Only one question remains: at today's price, is there any reason left to hold this trade?
- [Countermeasure] Delete "how much I'm already down" from the decision variables; answer only "at the current price, do my rules say buy or sell?" Ban averaging down once a position breaches its planned stop (any add-on counts as a new decision requiring the full entry process).
2.2 Disposition Effect (Disposition Effect)
- [What it is] Selling winners too early and holding losers too long — "sell gains, keep losses."
- [Trading example] Taking profit instantly at +5%, holding stubbornly at -15% "waiting to break even." Research suggests retail accounts broadly show this pattern: realized-sales lists contain far more winners than losers. The result is a high win rate (70%+) with a shrinking account — because average wins are far smaller than average losses.
- [Countermeasure] Hard-code exit rules: winners exit via trailing take-profits, losers via fixed stop — neither set by "feelings." "Feels about time" is not an exit condition. In reviews track "average win ÷ average loss," requiring ≥ 1.5; if the disposition effect drags it below, exits are still emotionally driven.
2.3 Endowment Effect (Endowment Effect)
- [What it is] Once you own something, you overvalue it. Same item, higher asking price than buying price; same stock, looks better and better once held.
- [Trading example] Held stocks are hard to sell whether they rise or fall: "my pick has great fundamentals." Research suggests people value their own holdings significantly higher than they valued them before buying. The "goodness" of your holdings is an illusion created by ownership.
- [Countermeasure] Once a quarter, re-evaluate every holding from a "flat perspective": if I had no positions and held cash instead, would I buy these at current prices? If not, sell.
2.4 Loss Aversion and Holding On (Loss Aversion)
- [What it is] The psychological pain of loss is roughly twice the pleasure of an equivalent gain (see article 01), so "realizing a loss" is nearly unbearable — easier to pretend it doesn't exist.
- [Trading example] Stop-losses placed, cancelled, re-placed; stop levels ratcheted lower and lower; "if I haven't sold, I haven't lost" — until a small loss becomes deep drawdown, and deep drawdown becomes zero.
- [Countermeasure] The only stop-loss form that works is "an exchange-resident stop order submitted simultaneously with entry" — let the system execute; leave yourself no window to cancel. Redefine "stop-loss" as "exiting per plan," not "admitting failure."
III. Self-Assessment Biases
3.1 Overconfidence (Overconfidence)
- [What it is] Systematically overestimating the accuracy of your own judgment and ability, especially after consecutive wins.
- [Trading example] After a few winning trades: adding leverage, increasing size, loosening take-profits — "I've got the touch." Research suggests accounts that trade more frequently and hold more aggressive positions tend to show worse net returns (evidence of negative correlation between overtrading and returns comes from data across multiple markets). A winning streak convinces you you're a genius; one deep drop is enough to prove otherwise.
- [Countermeasure] Position size is decided by formula, never by "touch." Enforce a cooling-off period after profits (stop trading for a few days once monthly profit exceeds a threshold). Track "share of predictions I got right" in your journal against actual win rate — the wider the gap, the harder you should rein in.
3.2 Self-Serving Attribution (Self-Serving Attribution Bias)
- [What it is] Crediting success to yourself (insight, judgment) and failure to externals (the market, market makers, news).
- [Trading example] Win: "my read was right." Loss: "shakeout," "terrible tape." Long term: good decisions and good luck both get logged as "I'm great," while bad decisions are systematically forgotten — your experience database is full of dirty data.
- [Countermeasure] Reviews evaluate only "did the trade comply with the rules," never profit or loss. Every trade must record "entry rationale" and "exit rationale" tied to specific rules, not "my judgment." Periodically reread original records of losing trades.
IV. Group & Probability Biases
4.1 Herding (Herding)
- [What it is] Abandoning your own judgment to follow the crowd — because blending into the group feels safest.
- [Trading example] A coin's 24-hour volume explodes, group chats flood, influencers shout calls, and fearing missing out you pile in — near the very top you become a provider of liquidity. "Everyone's buying, there must be something to it" is the shared faith of every bubble.
- [Countermeasure] Explicitly list "everyone is buying/selling" as a contrarian inspection signal. Before buying, ask: does my reason depend on "others are buying too"? If yes, cancel it. Study volume and turnover positioning in historical bubbles to build the intuition "crowded = dangerous."
4.2 Law of Small Numbers (Law of Small Numbers)
- [What it is] Treating "small samples" as "big laws": a handful of trades or moves, and suddenly there's a "pattern."
- [Trading example] "I bought five dips in a row and won every time — this strategy is solid." Five trades statistically prove nothing. "This coin pumps every time it breaks its prior high" — maybe just luck in a bull run. The "pattern" you see is often randomness handing you a self-deception.
- [Countermeasure] Any strategy needs dozens of valid samples before statistics mean anything (varies by strategy type). Validate "patterns" with small size; use written records, not memory. Stay suspicious of "worked the last few times": ask whether an economic mechanism supports the pattern, or if it's just numerical coincidence.
V. How Biases Stack Together
Each bias alone looks like "just a small flaw," but in trading they show up in teams and feed each other — that's the real mechanism behind chronic account bleed. The most common combinations:
5.1 The "Trapped Cycle": Anchoring + Loss Aversion + Sunk Cost
text
Buy → set a cost-based reference point (anchoring)
→ price falls → realizing the loss hurts too much, hold on (loss aversion)
→ keeps falling → "already down this much, cutting loses even more" (sunk cost)
→ bounces back near cost → "exit at breakeven," but it turns down before reaching cost (anchoring again)
→ loop until liquidation or deep entrapmentBreaking point: the weakest link in the chain is anchoring — delete "cost basis" from your decision variables and the other two biases lose their fuel.
5.2 The "Revenge-Trading Cycle": Loss Aversion + Reflection Effect + Overconfidence
text
Loss → "I must win it back" (loss aversion turns breakeven into obsession)
→ risk preference flips in the loss domain; willing to bet big (reflection effect)
→ occasional wins → "my method was right all along" (self-serving attribution + overconfidence)
→ double the size again → one deep drop gives it all backBreaking point: delete "make it back" from your dictionary. The market doesn't know you lost money; it only knows the current price — "getting back to even" is an obsession of mental accounting, not a trading goal (see article 04).
5.3 The "Self-Reinforcing Hype Loop": Confirmation Bias + Herding + Small Numbers
text
See everyone discussing a coin (herding) → buy
→ collect only bullish news (confirmation bias)
→ last few trades all won (law of small numbers) → "my read is sharp"
→ add size near the top → crashBreaking point: the bear-case list (the countermeasure for confirmation bias) also blocks herding — if you can't write 3 bearish reasons, you haven't thought independently at all; you're just following the crowd.
5.4 Combination Quick Reference
| Combination | Symptom | High-Risk Group |
|---|---|---|
| Anchoring + loss aversion + sunk cost | Trapped and holding ever longer | Beginners with positions |
| Loss aversion + reflection + overconfidence | Revenge trading after losses, losses snowball | People on losing streaks |
| Confirmation + herding + small numbers | Chasing hot themes ever higher | News-driven traders |
| Availability + hindsight | Entering on get-rich stories, feeling afterwards they "knew early" | Novices early in their journey |
| Endowment + disposition effect | Can't sell own picks, dump winners fast | Medium/long-term holders |
💡 Core Insight: Biases Are Interlocking Gears
Core insight: biases are not isolated failures but interlocking gears. Individual countermeasures help, but the truly effective defense is the trinity of "checklist + hard rules + review" (see article 05's anti-human-nature checklist) — cut any single link in the chain and the whole machine stops turning.
💀 Iron Law: Biases Are Not Isolated Failures but Interlocking Gears
Biases are not isolated failures but interlocking gears. Anchoring + loss aversion + sunk cost team up into the "trapped cycle"; loss aversion + reflection effect + overconfidence team up into the "revenge-trading cycle" — so single-bias fixes are not enough; you need the trinity of "checklist + hard rules + review" to cut a link in the chain and stop the machine.
VI. The Bias Self-Check List
Print it and stick it next to your monitor. Run through it before every order and after every win or loss.
✅ Conclusion: Run Through It Before Every Order and After Every P&L Event
Print it and stick it next to your monitor. Run through it before every order and after every win or loss. This 12-item checklist isn't theory you "read and remember" — it's an operations manual that belongs in your line of sight. Before orders, check "is my entry based on rules or feelings?"; after P&L, check "was this money skill or luck?" This is the most direct way to turn every cognitive bias from implicit to explicit.
text
[Before placing an order]
□ Is my entry reason based on rules, or on a "feeling"?
□ Have I written at least 3 reasons against this trade?
□ Am I tempted only because "everyone is buying/selling"?
□ Is my position size computed by formula, or manually bumped?
□ Am I trying to use this trade to "get back to even" or "prove myself"?
[While holding]
□ Am I in love with my cost basis? (anchoring)
□ Am I seeing only good news, blind to bad? (confirmation bias)
□ Am I holding just because "I'm already down too much"? (sunk cost)
□ Do I think my pick is "extra special"? (endowment effect)
□ Is the stop order still live on the exchange? Did I quietly cancel it?
[After a win or loss]
□ Won: was this money skill or luck? Would I add to the position? (overconfidence)
□ Lost: did I blame the market or my own decision? (self-serving attribution)
□ Did I rewrite what I believed at the time based on the outcome? (hindsight)
□ Was this trade "selling winners, keeping losers"? (disposition effect)
□ Does the "pattern" I just summarized have enough samples? (law of small numbers)VII. Master Quick-Reference Table
| Bias | One Sentence | Trading Scenario | Fastest Countermeasure |
|---|---|---|---|
| Anchoring | Hostage to the first number | Waiting to break even before selling; chasing off historical highs | Flat-position test; hide cost basis |
| Representativeness | Recent shape = permanent law | One month up, declare a bull market | Operational definition of trend |
| Availability | Memorable = probable | Heavy position on a get-rich story | Replace stories with statistics |
| Confirmation | Seeing only what you want | Bullish news amplified after entering, bearish ignored | Mandatory contra-reasons before orders |
| Hindsight | "Knew it all along" after the fact | "I should have sold earlier" | Record judgments at decision time |
| Sunk cost | What's spent hijacks the future | Averaging down while trapped | Ask only: at this price, should I act? |
| Disposition effect | Sell gains, keep losses | Average win 5%, average loss 15% | Rule-based exits for everything |
| Endowment effect | Owned = overvalued | Your own pick always looks best | Quarterly flat-perspective revaluation |
| Loss aversion | Losing hurts twice as much | Holding stubbornly, cancelling stops | Exchange-resident orders; no cancellation window |
| Overconfidence | Winning means "I'm brilliant" | Adding leverage after a streak | Position size by formula only |
| Self-serving attribution | Wins mine, losses the market's | "My read was right / it was a shakeout" | Review compliance only |
| Herding | Safest with the crowd | Chasing highs when group chats flood | "Everyone's buying" = contrarian signal |
| Law of small numbers | A few wins become a law | "Won five in a row" | No pattern without enough samples |
⚠️ Risk Warning
This content is for study and research only and does not constitute investment advice. A bias checklist is a self-awareness tool, not a trading signal: recognizing a bias doesn't guarantee avoiding it (between knowing and doing lies the entire discipline engineering of Chapter 07). If you find yourself looping through "knowing I should stop out but being unable to," return to Trading Psychology and let process and tools backstop your willpower.