Genius Doesn’t Protect You
The myth: smarter traders make better decisions. Two Nobel laureates. $124 billion in assets. $4.6 billion gone in 4 months.
Tuesday 9 June 2026 – Available on Spotify, Apple Podcasts, YouTube, Amazon Music
Why This Myth Exists
In almost every competitive field, intelligence is protective. Better analysis produces better outcomes. So when traders build sophisticated models and backtest rigorously, the intuition is that this makes them safer. It does up to a point. Past that point, the model becomes so convincing that the trader stops questioning whether the model itself could be wrong. They stop sizing for the scenario where their analysis is correct but the market does not care. This failure mode has a name: model risk.
- Model risk: why a correct strategy at the wrong size is indistinguishable from a wrong strategy
- VaR – what Value at Risk measures, what it misses, and why it fails in crises
- Correlation collapse: the mechanism that destroys diversified portfolios in a liquidity crisis
- The overconfidence trap: why high intelligence correlates with higher conviction and higher leverage
- Three rules for sizing positions when your model is strong
The Myth Destroyed – Model Risk and Leverage
Long-Term Capital Management ran convergence arbitrage – buying underpriced instruments and shorting overpriced ones in the same asset class, betting the spread would close. Individual trades were low-risk. Each spread had a small expected loss in the worst case. The problem was leverage and correlation assumptions.
With $4.7 billion equity supporting $124.5 billion in assets – 25:1 leverage – a 4% adverse move across the portfolio wipes equity completely. That sounds manageable until you understand what happens to correlations in a liquidity crisis.
VaR – The Measurement Tool That Failed
Value at Risk asks: given historical data, what is the maximum loss expected on X% of trading days? LTCM calibrated VaR to a 99% confidence interval using recent data. Two critical weaknesses emerge here.
It uses historical correlations. When Russia defaulted in August 1998, every asset class moved simultaneously. The correlations VaR was built on collapsed overnight. The model had never seen a global liquidity crisis and could not model one.
It ignores tail severity. VaR tells you the threshold below which losses stay on 99% of days. It says nothing about how bad the remaining 1% could be. LTCM lost $553 million in a single day – an event their model classified as essentially impossible. The probability estimate was wrong. The severity estimate was catastrophically wrong.
The Overconfidence Mechanism
Cognitive research is consistent: intelligence correlates with prediction accuracy up to a threshold, after which confidence keeps rising while accuracy plateaus. High-intelligence traders build more elaborate, internally consistent models – which feels like certainty but is actually conviction without calibration.
The pattern appears at every level. LTCM’s Nobel laureates trusted their models over the scenario analysis that would have revealed the leverage flaw. Isaac Newton correctly identified and exited the South Sea Bubble – then re-entered near the top because watching others profit without him was psychologically intolerable. Both failures were not analytical failures. They were humility failures.
Three Principles That Replace the Myth
“Not only did we underestimate the risks we were taking – we didn’t understand the nature of the risks.”
– John Meriwether, LTCM founder
“I can calculate the motions of heavenly bodies, but not the madness of people.”
– Isaac Newton, after the South Sea Bubble
Episode Timestamps
Continue Learning
- Risk of Ruin: The Mathematics Every Trader Must Understand
- Advanced Risk Management for Active Traders
- Position Sizing: The Complete Guide
The Complete Trader’s Edge
The Mind pillar covers the overconfidence trap, model risk, and building humility into your risk framework.
Educational purposes only. Not financial advice. Trading involves significant risk of loss.
