Overconfidence: 90% of Traders Think They’re Above Average

There’s a classic psychology survey: 90% of drivers believe their driving skills are above average. This is mathematically impossible—you can’t have 90% of people above the 50th percentile median. But in trading, this phenomenon is even more extreme: nearly all retail traders believe they can beat the market when they start, while statistics ruthlessly show that over 90% of retail traders ultimately lose money.

This is Overconfidence Bias—a universal human tendency to overestimate one’s abilities and knowledge. In trading, it’s the starting point of most catastrophic losses.

Core Principles

1. Better-Than-Average Effect: Systematic Overestimation of Self-Evaluation

The most classic manifestation of overconfidence is the Better-Than-Average Effect. People consistently rate themselves above objective levels across various ability dimensions:

• 90% of drivers think they drive better than average
• 80% of students think their academic ability is above average
• In trading, almost every beginner believes they can identify trends, control risk, and outperform the market

This overestimation isn’t built on real ability—it stems from a deep cognitive illusion: we’re confident about what we know, but almost oblivious to what we don’t know.

2. Illusion of Knowledge: More Information Means More Confidence (But Not More Accuracy)

An important source of overconfidence is the Illusion of Knowledge—when people receive more information, their confidence in their judgments increases, but judgment accuracy doesn’t necessarily improve proportionally.

In trading:

• You read 10 BTC analyses and your confidence in BTC’s direction surges
• But those 10 analyses might all say the same thing (bullish)—you just saw the same information 10 times
• The increase in information volume gives you the feeling of “I have a lot of information,” not “I understand the market more comprehensively”

Experiments prove: when professional horse bettors were given more information, their confidence in predictions rose significantly, but prediction accuracy barely improved. Information ≠ knowledge, and knowledge ≠ judgment.

3. Illusion of Control: Believing You Can Influence Random Events

Another core source of overconfidence is the Illusion of Control—people tend to believe they can influence outcomes that are fundamentally driven by randomness.

Typical manifestations in crypto:

• “I chose this coin, so it’ll go up”—actually the market has hundreds of variables
• “My analysis was spot on”—actually you might just have been lucky
• “I know when to enter and exit”—actually the market is highly uncertain

The illusion of control makes you treat luck as ability and randomness as controllability. When you’ve made money on several consecutive trades, the control illusion gets amplified dramatically—you feel you’ve “mastered the market’s rhythm,” when actually you may have just happened to be on the right side of a string of random fluctuations.

4. Self-Attribution Bias: Wins Go to Skill, Losses Go to Luck

Overconfidence is closely linked to self-attribution bias:

• Made money → “My analysis is impressive” → confidence expands → increase position size
• Lost money → “Market anomaly/manipulation/bad luck” → confidence remains intact → continue the same strategy

This asymmetric attribution lets confidence only rise, never fall. After 3 consecutive wins you think you’re a genius, but after 3 consecutive losses you don’t think you’re mediocre—you think it’s “the market’s fault.” This mechanism makes overconfidence snowball continuously.

5. Precision Overconfidence: Giving Overly Narrow Ranges

Overconfidence also manifests as Precision Overconfidence—people are too confident about the precision of their estimates. When asked to give a 90% confidence interval (the range where they believe the true value will fall with 90% probability), people’s intervals are usually too narrow, and actual hit rates are far below 90%.

In trading:

• You predict BTC will fluctuate between $95,000–$98,000 next week (you feel 90% certain)
• Actually BTC might swing to $88,000 or $105,000
• Your precision in price prediction far exceeds the information you actually have

Crypto Applications

Case 1: The Beginner’s “Three-Day Enlightenment” Illusion

A common phenomenon in crypto: beginners feel they “understand it” after just 3 days.

• Day 1: Bought BTC, it went up → “My judgment is accurate”
• Day 2: Bought an altcoin, it went up too → “I’ve already mastered the market rhythm”
• Day 3: Increase position size, feeling ready to make big money

This “three-day enlightenment” illusion is an extreme manifestation of overconfidence. What beginners gain in 3 days isn’t trading skill—it’s a few lucky random outcomes. But the illusion of control and self-attribution bias make them attribute all this luck entirely to “ability,” with confidence inflating to dangerous levels within just 3 days.

Statistics: Most crypto retail traders’ peak losses occur 2–4 weeks after entering the market—exactly the period when overconfidence has inflated and positions have been increased.

Case 2: Analysts’ Overconfidence Transmission

Crypto analysts’ overconfidence doesn’t just affect themselves—it infects their followers:

• An analyst posts “BTC will 100% break $150,000 within 3 months”
• Followers get infected by the analyst’s confidence, feeling safe “following a professional”
• Actually no one can predict the market with 100% certainty
• When the prediction fails, the analyst won’t admit error—they’ll say “the timeline shifted” or “the logic was right, just the timing was off”

This transmission of overconfidence forms a dangerous “confidence bubble”—analyst’s confidence → followers’ confidence → market sentiment’s confidence → eventual collapse.

Case 3: The Fatal Combination of Leverage and Overconfidence

Overconfidence’s most lethal manifestation is its combination with leverage:

• Overconfident traders feel their directional judgment is “almost certain”
• Because they’re “certain,” they think放大 leverage just “amplifies profit,” not “amplifies risk”
• 5x leverage → 10x → 20x → 50x
• Each leverage increase is overconfidence pushing the wave higher

Data from CEXs: Among retail traders using 10x+ leverage, 95% get liquidated within 30 days. Overconfidence makes them feel “I can handle high leverage,” while reality says no one can sustainably handle high leverage.

Practical Scenarios

Scenario 1: Confidence Calibration Training

Build a confidence calibration log:

Date	Prediction	Confidence	Actual Result	Correct?
7/1	BTC up this week	80%	Went up	✓
7/8	ETH down this week	70%	Went up	✗
7/15	SOL breaks 200	60%	Didn’t break	✗

After recording for 30 consecutive days, calculate: what’s the actual hit rate of predictions you marked as 80% confidence? If it’s only 50%, your definition of 80% confidence is too loose—you need to systematically lower your confidence calibration.

Calibration goal: predictions marked as 70% confidence should have a 70% hit rate. If the actual hit rate is lower than the calibrated confidence, you’re overconfident.

Scenario 2: Position Size Linked to Confidence

The most direct method against overconfidence: make position size strictly linked to calibrated confidence, not to emotional confidence.

• Confidence ≤ 50%: Don’t trade
• Confidence 50–60%: 1% position
• Confidence 60–70%: 2% position
• Confidence 70–80%: 3% position
• Confidence ≥ 80%: 5% position (rarely occurs)

Note: Even when you feel “90% certain,” use only 5% position. Because your calibration records likely tell you—your 90% confidence predictions probably hit at only 60%.

Scenario 3: Forced “Devil’s Advocate Report”

Before every major trading decision, write a “devil’s advocate report”—pretend you’re someone arguing against your own decision, listing 3–5 counterarguments. If the devil’s advocate report makes you think “that actually makes sense,” your confidence should be lowered. If you think the counterarguments are “completely untenable,” you can maintain higher confidence—but remember, this itself may also be influenced by confirmation bias.

Common Misapplications

Misapplication 1: Equating “Confidence” with “Overconfidence”

Reasonable confidence and overconfidence are different. If you have sufficient evidence, calibrated judgment, and sound risk management, then moderate confidence is normal. Overconfidence means confidence levels far exceed actual ability—the difference lies in whether there’s objective calibration data to support it.

Misapplication 2: Overcorrecting into “Complete Uncertainty”

Some people, after realizing overconfidence, become “uncertain” about all judgments—refusing to open any position or make any decision. This isn’t overcoming overconfidence—it’s replacing one extreme with another. The correct method is calibration: making confidence match actual hit rates, not eliminating confidence itself.

Misapplication 3: Thinking Only Beginners Have Overconfidence

Overconfidence isn’t exclusive to beginners. Experienced traders, fund managers, and even Nobel laureates in economics have documented cases of overconfidence. Experience may help reduce certain types of overconfidence, but it can also create the meta-bias of “I’m experienced so I won’t be overconfident”—overconfidence itself makes you believe you’ve already overcome overconfidence.

Misapplication 4: Using Past Returns to Prove You’re “Not Overconfident”

“I’ve made 200% over 3 years, so my confidence is justified.” But 3 years of returns may just be luck—market cycles, industry trends, or specific coin explosions can give you high returns unrelated to ability. You need longer calibration data and stricter statistical tests to distinguish ability from luck.

Summary

Overconfidence is the most隐蔽 and dangerous cognitive bias in trading. It doesn’t make your heart race like fear, or make you impulsively place orders like greed—it quietly makes you overestimate every judgment, unconsciously increasing positions, narrowing stop losses, and ignoring risk.

There’s no shortcut against overconfidence, only one path: Calibration. Record every prediction and confidence level, compare against actual results, find the real gap between your confidence and hit rates, then adjust decisions based on calibration data. This isn’t about making you less confident—it’s about making you accurate. Accurate confidence is real confidence; inflated confidence is just an illusion.

In crypto, when you feel “this time I’m definitely right,” pull out your calibration log—when you felt “definitely right” historically, how many times were you actually right? The numbers might cool you down.

For more practical methods, see Dimen Trading.