cryptotradingbook

Cognitive biases

Foundations → Intermediate · ~12 min read

At a glance

  • Your brain runs on shortcuts that were survival tools - and are trading liabilities.
  • Loss aversion: losses hurt roughly twice as much as equal gains feel good.
  • Confirmation bias reads the chart as a mirror; sunk-cost keeps dead positions alive.
  • Overconfidence after wins is the standard prelude to oversizing.
  • Every bias is countered by a written rule, not by willpower.

1. Why your brain is the risk

Trading punishes precisely the intuitions that served evolution: fast pattern-matching, loss avoidance, social imitation. The market is adversarial and random at the timescales that feel meaningful - the worst combination for a brain built for weather and tribes. Knowing the biases is not the fix; rules are the fix, because rules run when the brain is offline.

2. The four that cost the most

Loss aversion - losses hurt ~2x gains feel good; the visible result is stop-avoidance and holding losers "until it comes back". Confirmation bias - after entry you read the market as agreement; your invalidation rule must be written before the trade for exactly this reason. Sunk-cost - "I'm already down 60%" is not a reason; it is a description of the past. Overconfidence - three wins in a row and size doubles; the position size formula is the antidote because it does not care how you feel.

Remember: You cannot out-willpower a bias. You can only pre-commit: the rule written when calm runs when you are not.

3. Bias in the wild: three portraits

The averaging-down without a plan (sunk-cost + loss aversion): each dip bought "cheaper" with no invalidation level - the position becomes a faith. The revenge trader (loss aversion inverted): a loss is "unacceptable", so size doubles to win it back fast - the single fastest account-destruction pattern there is. The prophet (confirmation + overconfidence): a winning call becomes an identity, and identities cannot stop out.

4. The countermeasure architecture

Every bias maps to one artefact: loss aversion → the stop in the plan; confirmation → the invalidation rule written pre-entry; sunk-cost → the exit rule and the journal's adherence column; overconfidence → the sizing formula. The system is boring on purpose: your best trading happens when your personality is least involved.

THE BIASTHE ARTEFACT (written, not willed)Loss aversionThe stop in the planConfirmation biasInvalidation rule, written pre-entrySunk-cost fallacyExit rule + journal adherence columnOverconfidenceThe position size formula

You cannot out-willpower a bias. You can only pre-commit: the rule written when calm runs when you are not.

Figure 20.1 — The countermeasure architecture: every bias maps to one artefact. The system is boring on purpose.

5. When this fails

The rules fail when they exist only in the head, when they are edited mid-trade, or when the journal starts grading feelings instead of adherence. A bias operates most freely exactly when you are sure you are above it - the professional marker is not the absence of bias but the presence of checks.

Risk: The most expensive sentence in trading is "this time I'm sure". Certainty and position size should have zero correlation - the sizing formula enforces that; nothing else does.

Knowledge check

  1. What is the actual fix for cognitive biases? (Written rules pre-committed when calm - not willpower.)
  2. Losses hurt about how much more than equal gains feel good? (Roughly twice - loss aversion.)
  3. Why must invalidation rules be written before entry? (Confirmation bias rewrites the reading after entry.)
  4. What is the fastest account-destruction pattern? (Revenge sizing after a loss - doubling to win it back.)
  5. Map the four biases to their artefacts. (Loss aversion→stop; confirmation→pre-written invalidation; sunk-cost→exit rule + journal; overconfidence→sizing formula.)