Why Do Past Winners Often Become the Biggest Losers?

A trader makes 40% returns in year one and believes they've discovered a skill. They make 30% returns in year two and assume their methods are proven. They increase position size in year three and face a 50% drawdown. They panic, exit at the worst time, and lose 70% of capital. This arc—from confidence to catastrophe—follows a psychological pattern called overconfidence bias. In financial markets, overconfidence is measured as the gap between how well you think you'll perform and how well you actually perform. Research shows that 93% of traders believe they're above-average (a mathematical impossibility; the average cannot be pulled upward), and traders who've experienced success are most at risk of overestimating their skill in tail events. This chapter examines overconfidence as a risk factor independent of leverage or correlation breakdown, explaining how it creates blind spots that force a reckoning.

Quick definition

Overconfidence trading risk is the gap between perceived skill (how well a trader believes they'll perform) and actual skill (how well they perform, especially in unseen tail events). Overconfidence causes traders to take larger positions, use more leverage, and refuse to accept losses because they believe their "system" will recover. When the system fails, losses are amplified by overconfidence-driven risk-taking.

Key takeaways

• Traders who've been profitable for 1–3 years are most at risk of overconfidence; they've never experienced a major drawdown and believe drawdown-level declines won't happen to them.
• Survivorship bias creates false confidence; a trader sees only the winners (successful strategies that survived) and doesn't see the failures (strategies that blew up, archived, and disappeared from discussion).
• Illusory skill confuses luck with ability; a trader who made 40% returns may have been 80% lucky and 20% skill. Year 3 exposes the luck, and the trader is surprised because they conflated luck and skill.
• Overconfidence causes traders to ignore warning signs: a position that falls 10% when the model predicted a 5% downside max is evidence that the model is wrong, but an overconfident trader sees it as a "buying opportunity" and adds size.
• The illusion of control is the most dangerous form of overconfidence; a trader believes they can always exit a position, predict a crash, or recover from losses. They can't. Markets don't care about the trader's intentions.

The psychology of overconfidence: Why winners are most at risk

Overconfidence is not stupidity; it's a normal feature of human psychology that is amplified in environments with delayed feedback, like trading. A surgeon receives immediate feedback (patient dies or recovers). A trader receives delayed feedback (trade result after days/weeks). The delayed feedback creates a window where the trader's confidence exceeds actual skill, and the bigger the early success, the wider the window.

A retail trader backtests a strategy, sees 40% annual returns in a 10-year test, and thinks: "This is reliable. I'm skilled." The backtest is usually right about the past (it's running on actual historical data) but wrong about the future (it tested only one historical period, not the next one). The trader doesn't notice the difference because the feedback is delayed. They live in the backtest world (40% returns, 8% drawdown) and assume the real world will match.

Winners are most at risk because success is self-reinforcing in confidence. A trader who made 40% in year one thinks: "I'm talented." A trader who made 40% and 30% in years one and two thinks: "I'm proven talented." A trader who made 40%, 30%, and 25% for three consecutive years thinks: "I'm so talented I can handle 4:1 leverage." Year four's 50% drawdown shocks them because their confidence has no experience parameter; the confidence is absolute, not relative to time-tested outcomes.

Survivorship bias: Seeing winners, not losers

A retail trader reads about the top 100 traders and their strategies on Reddit or Twitter. All 100 are profitable. The trader assumes that skilled trading is possible and attempts to copy the strategies. They don't see the 10,000 traders who tried similar strategies and failed; those traders are silent (they quit, are embarrassed, or are in debt). The visible set is survivorship-biased toward winners.

This creates a false estimate of skill availability. A trader might think: "If 100 people are profitable, the base rate of profit-making is 100/100 = 100%." In reality, the base rate might be 100/10,100 = 0.99%. The denominator is hidden.

Survivorship bias is amplified in bull markets. A bull market (2015–2021, for example) makes most traders profitable. A trader experiencing 15+ years of bull market assumes market skill is common. They're wrong; the bull market was doing the heavy lifting. A trader tested in a bear market would fail.

The 2008 financial crisis exposed this. Traders and managers who were profitable for 15 years straight (2000–2015 bull market) suddenly blew up when the market declined. They were not skilled traders; they were beneficiaries of a favorable regime. The regime change revealed the lack of skill.

Illusory skill: Luck vs. ability in trading

A trader's 40% annual return over three years might come from:

1. 20% skill (good strategy, good execution)
2. 70% luck (favorable market regime, tail winds)
3. 10% leverage (amplifying returns by using borrowed money)

The trader perceives it all as skill, so they:

1. Increase position size (assuming the skill will continue)
2. Leverage deeper (assuming the advantage is real)
3. Diversify into new markets (assuming the skill is transferable)

When market regime changes (luck disappears), the strategy collapses. The trader is shocked because they thought the 40% was repeatable skill, not a confluence of luck and regime.

How does a trader distinguish skill from luck? Through diversification and time. A strategy that works in multiple markets, multiple regimes, and multiple time periods is more likely to be skill. A strategy that works only in one favorable period is likely luck. But overconfidence prevents this distinction. A lucky trader feels certain their strategy is skilled, so they don't diversify; they double down.

Case 1: The trader who made 60% in year one, then blew up

A retail trader on Interactive Brokers opened an account with $50,000 in January 2020. Using a mean-reversion strategy (buying oversold stocks, selling overbought stocks), the trader made $30,000 in 10 months (a 60% return), riding the post-COVID market recovery. The trader was convinced they'd found the edge.

In November 2020, the trader quit his job and decided to "go full-time." He leveraged the account to $150,000 using margin (3:1 leverage). His strategy required buying dips and holding for 3–5 days until the stock rebounded. The strategy worked 15 times in a row (March–October 2020).

In December 2020, the Fed began discussing rate hikes. The market fell 2% in a week. The trader's mean-reversion strategy should have shorted "overbought" technicals, but the Fed signals were too strong. The market fell another 5%. The trader's positions reversed direction. He had losses on 7 out of 10 positions. He doubled down on the two strongest positions, adding leverage (now 4:1).

By January 2021, the market had fallen 8% total. The trader's $150,000 leveraged position had lost $40,000 (a 27% loss on equity). His equity was now $60,000, and his margin requirement was 30% of position value, or $45,000 (30% of $150,000). His cushion was $15,000. The broker issued a margin call warning: "Additional losses of 10% will trigger forced liquidation."

The trader was faced with a choice: accept that his "edge" didn't work in all market conditions, reduce leverage, and move on with a smaller loss; or hold and hope the market reverted. He held. He was overconfident in his strategy; he believed the mean-reversion signal was right and the market was wrong.

The market fell another 8% before the Fed pivoted and the market recovered. During the downside, the trader's margin was liquidated at $8,500 loss on top of the $40,000 unrealized loss. Total damage: $48,500, a 97% loss of the original $50,000. The trader's job was gone, his account was gone, and he'd borrowed an additional $10,000 on credit cards trying to average down in the final week.

Case 2: The options seller who thought he had zero risk

A trader on Tastyworks had a profitable options-selling strategy for 18 months: sell cash-secured puts (on stocks with 20% downside protection), collect premiums, and let them expire worthless. He averaged $3,000 per month in profit for 18 months, a total of $54,000 on a $100,000 account. The account was up 54%, and he believed the strategy was "low-risk" (the puts required $50,000 in cash collateral, so "only" 50% of the account was at risk).

In month 19, his short puts were tested. A Fintech stock he'd sold puts on fell 25% overnight on bad earnings. His short $200 puts were now $195 in-the-money (intrinsic value + time value). His max loss was $10,000 per contract; he'd sold 5 contracts, so his max loss was $50,000 (half his account).

He faced a choice: close the puts at a $40,000 loss and accept the failure, or hold to expiry hoping the stock would recover. He held. He believed the stock was "a great company" and "temporary weakness." He used his conviction (overconfidence) to override the math.

The stock fell further. At expiry, it closed at $165 (a 35% decline). He was assigned stock at the $200 strike. He now owned 500 shares at $200 each ($100,000 cost) but the shares were worth only $82,500 in the market. He had a $17,500 embedded loss on top of the $50,000 loss from the options themselves.

But he couldn't sell the shares immediately; he was out of buying power (all capital was used to pay for the shares). He had to wait 2 trading days to sell and process the sale. By then, the stock had fallen another 3%. He sold at $160 per share, realizing the full $50,000 + $20,000 = $70,000 loss, or 70% of his original $100,000 account.

The most damaging part: he believed the strategy was "low-risk" because he'd collected $54,000 in premium over 18 months. The premium made him think the risk was asymmetric in his favor. In reality, option-selling is short-volatility, and volatility can spike and erase months of premiums in a single event.

Case 3: The crypto trader who thought they understood the market

A trader who'd profited from Bitcoin and Ethereum purchases in 2021–2022 decided they understood crypto "better than most." They began a Twitter account, shared analysis, and built a following of 50,000 people. The attention fed overconfidence.

In 2023, they used 5:1 leverage to short Ethereum, believing it was "overvalued" at $2,200. The trade was based on their proprietary analysis, and they were certain they were right. They posted on Twitter: "I'm short ETH at $2,200. The smart money is exiting. Watch for capitulation."

Ethereum rose from $2,200 to $2,500 (+14%). Their leveraged short lost $70,000 on a $100,000 position (a 70% loss of equity). The leverage meant forced liquidation was imminent. Instead of closing the position and accepting a $70,000 loss, the trader doubled down, posting: "Institutions are accumulating before the rally. I'm adding to my short."

Ethereum continued to $2,800. The trader's position was liquidated automatically by the exchange at $2,780. They lost the remaining $30,000 of equity and owed the exchange $40,000 for the forced liquidation slippage and fees.

The most damaging part: the trader's 50,000 Twitter followers had followed the trade. Dozens had copied the short. When the trade failed, the followers also lost money. The trader had passed their overconfidence on to others.

The illusion of control: Believing you can always exit

Traders often believe they can exit a position whenever they want. This illusion of control is especially dangerous in futures, options, and crypto, where execution risk is high. A trader shorting illiquid altcoin with 10:1 leverage believes: "If the price rises 10%, I'll just close the position." But the altcoin has $10 million daily volume, and the trader's $1 million position is 10% of daily volume. Closing takes days, and slippage is 5–10%.

The illusion of control also affects position sizing. A trader believes: "I can handle a 30% position drawdown because I can always sell." But "always" is an illusion. During a market panic (March 2020, September 2023 bank crisis), sell orders pile up, bid-ask spreads widen, and "always selling" becomes "hopefully selling at some price."

A trader using 10:1 leverage believes: "If the position drops 10%, I'll get margin-called and forced-liquidated, but I can manage that." In reality, margin calls happen during the hours when markets are most volatile and execution is worst. The trader can't "manage" it because the forced liquidation is nonoptional.

Overconfidence after a streak of success

The most dangerous period for a trader is 6–18 months after a major success. A trader who made 100% returns in the previous year believes they'll make 50%+ the next year. They don't plan for a drawdown because they've internalized the success as "repeatable." When the drawdown comes (and it always does), they're shocked and unprepared.

Research by psychologist Daniel Kahneman shows that people's confidence in their ability increases linearly with success, but their actual ability remains roughly flat. A trader who's won 10 trades in a row is more confident than a trader who's won 3 in a row, but they're not actually more skilled. The 10-win streak is 80% likely due to luck (volatility, favorable regime) and 20% likely due to skill.

The specific overconfidence trap of backtesting

Traders backtest a strategy on 10 years of historical data and observe 15% annual returns with a maximum 20% drawdown. They launch the strategy with real money. For the first 2 years, the strategy works: 13% and 18% returns. The trader believes the backtest was accurate and that they've simply been lucky (getting good returns).

In year 3, the strategy faces a 35% drawdown (worse than any drawdown in the backtest). The trader is shocked because they thought the maximum drawdown was 20%. The problem: the backtest was run on one 10-year period. The 20% maximum drawdown was the largest drawdown that occurred during 1990–2000 (for example). The drawdowns of 2008, 2020, or 2022 were not in the backtest period, so the strategy was never tested against them.

Overconfidence based on backtest results is the most common form of overconfidence among retail traders.

How overconfidence interacts with leverage

Overconfidence + leverage is the worst combination. A trader who's confident in a strategy uses 2:1 leverage to amplify returns (confident they're right). When the strategy is wrong (a drawdown occurs), the leverage amplifies losses. A 20% drawdown on 2:1 leverage is a 40% loss of equity. The trader is underwater fast, and forced liquidation comes sooner than expected.

A trader without overconfidence might size the position at 1:1 (no leverage) and tolerate a 20% drawdown as an acceptable risk. The same trader with overconfidence sizes at 2:1 and watches a 20% drawdown become a 40% equity loss, a margin call, and forced liquidation.

Overconfidence is the psychological factor that causes traders to over-leverage. Without overconfidence, most traders would recognize that 25:1 leverage on bond arbitrage (LTCM) or 10:1 leverage on crypto (FTX, Three Arrows) is dangerous. With overconfidence, they believe they have an "edge" that justifies the leverage.

Real-world examples of overconfidence blowups

Amaranth Advisors (2006): A $9 billion hedge fund focused on energy trading made 15% returns for three years. The fund's traders became overconfident in their ability to trade natural gas futures. They took on a massive concentrated position in energy derivatives, assuming their skill would protect them. A sudden market shift (mild winter reducing natural gas demand) caused a $6 billion loss in a single week. The fund's overconfidence in energy-trading skill was not matched by actual diversification or risk management.

SAC Capital (2013): A $15 billion hedge fund with a long track record of 20%+ annual returns became overconfident that their insider-trading network was "legal" and "safe." The fund faced SEC charges, paid $1.8 billion in fines, and eventually shut down. The overconfidence was that the edge they'd identified (information asymmetry) was sustainable and legal. It wasn't.

Robinhood traders (2020–2021): Retail traders who profited during the 2020–2021 bull market became overconfident that they could daytrade. Many used margin to amplify returns, trading options, meme stocks, and other volatile instruments. A decline in the market revealed that their prior returns were driven 90% by the bull market and 10% by skill. Positions were liquidated, and accounts were wiped out.

Common mistakes

1. Assuming three years of profits prove skill: Three years of profitability in a bull market proves nothing about skill. Test your strategy in a bear market, a sideways market, and a regime-shift market. If it fails in any of these, it's not robust.

2. Increasing leverage after a winning period: The worst time to leverage higher is when you've just been profitable. That's when overconfidence is highest and humility is lowest. Decrease leverage after wins; increase it after losses (while reducing position size).

3. Ignoring warning signs that the strategy has changed: A strategy that worked for three years but is now flat for six months is a warning sign that the market regime has shifted. An overconfident trader assumes the shift is temporary and doubles down. The strategy is often permanently broken in the new regime.

4. Dismissing losses as "noise" or "bad luck": A 15% loss when the strategy's historical drawdown is 10% is evidence that the strategy is wrong in the current market. Dismissing it as noise and holding is overconfidence.

5. Believing your edge is timeless: The edges you discovered 3 years ago (maybe mean-reversion on small-cap stocks, or technical patterns on 4-hour charts) might be crowded now, beaten down by competition, or broken by new market structure. Assume your edge has a shelf life; periodically test whether it still works.

FAQ

How do I know if I'm overconfident?

Ask: "What's the worst outcome that could happen to this trade, and am I prepared for it?" If your answer is vague (e.g., "It's a great company, it won't fall 50%") or dismissive ("That's very unlikely"), you're overconfident. If your answer is specific and you've stress-tested it (e.g., "If it falls 30%, I will lose $X and I'm comfortable with that loss"), you're managing risk correctly.

Can I reduce overconfidence?

Partially. Mechanical rules help: enforce stops, enforce position-sizing limits, enforce leverage limits. A rule like "Never use more than 1.5:1 leverage" removes the decision from overconfident hands. Annual strategy audits also help: review your assumptions, test your strategy on unseen data, and acknowledge when it's no longer working.

Why is overconfidence worse than other trading mistakes?

Because overconfidence causes you to compound other mistakes. An overconfident trader who uses leverage will size positions too large. An overconfident trader who backtests will backtest only favorable periods. An overconfident trader who faces a loss will hold instead of accepting it. Overconfidence amplifies all other mistakes.

Is overconfidence the same as the Dunning-Kruger effect?

Related but not identical. The Dunning-Kruger effect is the tendency of low-skill people to overestimate their skill. Overconfidence is the tendency of all people (low-skill and high-skill) to overestimate their ability to predict the future. A skilled trader can be overconfident about their ability to predict tail events they've never seen.

Do professional traders have less overconfidence than retail traders?

No. Professional traders work in environments with large amounts of capital and peer feedback, which can reduce overconfidence. But many professional traders and hedge fund managers are just as overconfident as retail traders (LTCM, Archegos, SAC Capital). The profession doesn't cure the bias.

Can past success ever be reliable evidence of future skill?

Yes, but only if the past success is tested across multiple regimes. A strategy that works in bull markets, bear markets, sideways markets, and crisis markets is more likely to be skill-based. A strategy that works only in one favorable regime is likely luck. Require evidence of skill across multiple regimes before increasing leverage or position size.

Related concepts

• Common Patterns Across Every Risk Disaster
• Leverage: The Common Thread in Every Blowup
• Retail Blowups: Margin Calls Gone Wrong
• The Risk of Ruin: When Your Account Hits Zero
• What Risk Actually Means in Markets

Summary

Overconfidence is a psychological risk factor that causes traders to take larger positions, use more leverage, and refuse to accept losses because they believe their "skill" or "system" will recover. Traders who've been profitable for 1–3 years are most at risk; they've never experienced a major drawdown and believe drawdowns won't happen to them. Survivorship bias (seeing only winners, not losers) creates false estimates of skill prevalence. Illusory skill confuses luck with ability; a 40% return might be 20% skill and 70% luck (favorable regime), but the trader perceives it all as skill. The illusion of control causes traders to believe they can always exit a position, manage a margin call, or tolerate leverage safely. When market regime shifts and the "untested" tail event arrives (a drawdown worse than any in the trader's experience), overconfidence is revealed. The trader faces forced liquidation, catastrophic losses, and ruin. The defense is mechanical: enforce stops, enforce position-sizing rules, enforce leverage limits, test strategies across multiple regimes (including bear markets and crisis periods), and assume past success is luck until proven skill by testing in unseen conditions.

Next

→ Due Diligence Lessons from Every Case Study