Dealing with Overconfidence
Signs & solutions
Overconfidence is probably the most prevalent cognitive bias among investors and entrepreneurs. I strongly believe that once we overcome this, the path toward +EV decisions and actions become much easier.
Signs of overconfidence
Let’s start with the symptoms.
Here are the clearest signs that an investor, trader, or entrepreneur is suffering from overconfidence, a bias that quietly depletes cognitive capital by blocking calibration, Bayesian updating, and respect for general base rates. In the framework outlined in Cognitive Capital, overconfidence isn’t mere arrogance; it’s a structural failure in probabilistic reasoning that replaces evidence-based edge with manufactured certainty. Your mind begins to treat uncertainty as a flaw to be eliminated rather than a feature to be acknowledged, leading to -EV (poor expected-value) decisions and eventual erosion of returns or venture survival.
Binary predictions over probability distributions
The overconfident operator speaks in absolutes, ”This stock will double,” “The market will crash next quarter,” “My startup will hit product-market fit in six months”, instead of assigning ranges or distributions (e.g., 60% chance of 20-40% return, 25% chance of flat, 15% downside). This illusion of precision avoids calibration training and invites catastrophic belief updates when reality diverges. In trading, it manifests as oversized bets without regard for full outcome potentials; in entrepreneurship, it fuels the refusal to pivot despite mounting contrary evidence.
Ignoring or dismissing base rates in favor of personal narrative
Base rates, the historical, generalized, default frequency of similar outcomes, are the anchor of sound probabilistic thinking. Yet the overconfident mind discards them for a seductive story: “This time is different because of my unique insight,” or “I’m so much smarter than the average founder.” Short term traders overweight recent wins while ignoring VERY NEGATIVE Expected Performance of average speculators; entrepreneurs convince themselves their idea is the exception to the 90%+ startup failure base rate. The result is systematically overestimated probabilities of success and underestimation of risk.
Excessive trading/management-adjustment frequency
Overconfidence expresses itself in hyperactivity: frequent investor entries and exits, business development (marketing, inventory, debt) adjustments beyond what evidence justifies. Studies in behavioral finance consistently show overconfident investors trade more often, incur higher transaction costs, and underperform. The underlying mechanism is illusion of control, the belief that personal skill dominates market randomness, preventing the reflective pause that good cognitive capital demands before any action.
Failure to update beliefs proportionally (Bayesian malpractice)
New evidence arrives, but the overconfident mind either ignores it (if disconfirming) or overreacts dramatically (if confirming). Proper updating requires proportional adjustment toward the prior (base rate + new signal). This rigidity destroys the belief-updating habit that compounds growth over time.
Narrative seduction overriding evidence
Stories are powerful; they reduce complex systems to clean, simple causal arcs. The overconfident operator falls for their own narrative, ”The Fed will pivot because they always do when markets wobble”, and suppresses base rates or disconfirming data that don’t fit. In investing, this leads to chasing momentum without probabilistic discipline; in entrepreneurship, it produces founder delusion where vision trumps market feedback.
Lack of decision journaling or post-mortem calibration
True calibration requires tracking forecasts against outcomes over many decisions, ideally via Brier scores or simple logs. The overconfident rarely do this; they attribute wins to skill and losses to external noise (self-attribution bias), so they never receive the feedback loop that would expose their miscalibration.
Motivated reasoning and confusion of probability with desirability
When the desired outcome is high (a moonshot trade, a unicorn exit), the mind inflates its probability to reduce internal dissonance. “I want this to work, so it probably will” replaces dispassionate estimation. This is especially lethal for entrepreneurs, where personal identity fuses with the venture, leading to persistent overestimation even after years of evidence.
These signs rarely appear in isolation; they often reinforce each other, forming a dangerous feedback loop that accelerates depletion of the very mental tools, reflective thinking, calibration needed for +EV mental models.
The Antidote
The antidote is deliberate practice: force base rates into every assessment, decompose questions, assign explicit distributions, journal rigorously, and treat overconfidence not as a personality trait but as a measurable, correctable error in reasoning.
There is no one-time fix; it’s a compounding loop of reflection, tracking, and constraint. Below is a list of to-do items I believe to be the most important. Implement them sequentially or in parallel, and track your performance for at least 90 days before evaluating progress.
Mandate base-rate anchoring on every major decision
Before any thesis, trade, or business commitment, write down the relevant historical base rate first (e.g., active crypto traders lose 36%/year on average, only 50% of cafes survive the first 5 years, annual ROI of bars are at -5%~ on average). Only then adjust with your private signal. Use a simple template: “Base rate: X%. My edge: Y%. Updated probability: Z%.” This single habit kills narrative seduction faster than anything else.
Decompose every question into explicit probability distributions
Replace binary statements (“This will work”) with ranges: “60% chance of 20-40% return, 25% chance of 0-10%, 15% chance of -30%+.” Do this in writing for every position >5% of portfolio or every major product/strategic pivot.
Maintain a decision journal with probabilistic forecasts
For every non-trivial bet (trade entry/exit, hiring decision, fundraising round, feature prioritization), log:
Date and decision
Explicit probability distribution at the time
Rationale (including base rates)
Outcome (resolved or still open) Review quarterly. This creates the feedback loop that overconfidence avoids.
Calculate and track personal Brier score or calibration curve
After 20–50 resolved forecasts, compute your Brier score (lower is better; perfect calibration = 0). Plot a calibration curve (actual frequency vs. predicted probability bins). Tools like PredictionBook, Good Judgment Open, or a simple spreadsheet suffice. Seeing miscalibration in data is far more convincing than introspection.
Run pre-mortems on every significant commitment
Before executing, write a short paragraph assuming failure: “It is 12 months from now and this trade/startup has blown up. What were the most likely causes?” List 3–5 realistic paths. This has helped me overcome overconfidence in a very big way.
Over time, these steps can convert overconfidence from a hidden tax on performance into a visible, correctable input. The key metric is not feeling humble, it’s seeing your actual outcomes increasingly match your expected distributions. That’s cognitive capital compounding at work.