The Psychology of Losing Streaks: Why Investors Abandon Good Strategies
Behavioral finance explains why drawdowns feel personal, why streaks look longer than they are, and how disciplined investors decide when to stay the course versus when to walk away.
The hardest part of investing is not finding a strategy. It is surviving the period when the strategy looks wrong. That is when investors start rewriting the story: the edge is gone, the market has changed, the model is broken, everyone else is smarter. Sometimes that is true. Often it is just a losing streak doing what losing streaks do—testing patience, identity, and discipline at the same time.
Behavioral finance has spent decades explaining why this happens. Kahneman and Tversky showed that people evaluate outcomes relative to a reference point and dislike losses more than they like equivalent gains [1]. Odean found that individual investors systematically realize gains too quickly and hold losses too long, a pattern now known as the disposition effect [2]. Dalbar’s long-running investor behavior studies have repeatedly shown that the average investor’s realized returns lag broad market benchmarks, largely because of timing mistakes and behavior under stress [3].
That combination—loss aversion, disposition, and recency—creates a predictable failure mode. Investors do not merely dislike drawdowns. They start to interpret them as evidence that the strategy itself has become invalid. That is the central psychological trap this article is about.
A losing-streak probability requires an explicit model
A run feels diagnostic because recent outcomes are salient, but length alone does not identify whether an edge changed. Calculate a base rate before interpreting it. The table uses independent Bernoulli trials with a constant loss probability and asks for at least one run of five consecutive losses in 100 trades.
These exact model probabilities are not forecasts for a trading strategy. Real returns can be serially correlated, non-binary, regime-dependent, differently sized, and changed by costs. Estimate those features from point-in-time data and stress worse dependence. Do not infer positive expectancy from win rate alone.
Probability of at least one five-loss run in 100 independent trades
Win rate
Loss rate
One fixed five-trade window
At least one in 100 trades
40%
60%
7.78%
97.58%
50%
50%
3.13%
81.01%
55%
45%
1.85%
64.68%
60%
40%
1.02%
45.91%
65%
35%
0.53%
28.52%
Loss aversion: the asymmetry that changes behavior
Loss aversion is the foundational idea here. In prospect theory, losses loom larger than gains of the same size [1]. That asymmetry is not a minor preference; it changes decision-making. Investors become more sensitive to recent drawdowns than to the long-run payoff distribution that justified the strategy in the first place.
This is why a systematic trader can know, intellectually, that a 15% drawdown is within historical norms and still feel compelled to intervene. The emotional system is not asking whether the strategy has positive expectancy. It is asking whether the pain can stop now.
The practical implication is uncomfortable but useful: if you do not pre-define the conditions under which you will change a strategy, your brain will define them for you in the middle of the drawdown. That is usually the worst possible time to make a structural decision.
Situation
Emotional interpretation
Process-based interpretation
Three losses in a row
The edge is gone
A short streak is expected in any noisy process
Strategy underperforms for one quarter
The market has rejected it
Short windows are too small to infer structural failure
A winner turns into a loser
I should have sold earlier
Outcome regret is not the same as process error
Drawdown exceeds comfort level
This is broken
This may be a risk-budget issue, not a signal issue
Comparison: what the brain hears versus what the strategy says
The disposition effect: why investors sell the wrong thing
Odean’s classic study of brokerage accounts found that investors were more likely to sell winning positions than losing ones, consistent with the disposition effect [2]. The logic is familiar: realize the gain to feel smart, defer the loss to avoid admitting error. But that behavior is costly. It can turn a disciplined process into a collection of emotional exceptions.
The disposition effect matters during losing streaks because it creates a double distortion. Investors cut the positions that are working and keep the ones that are not. Then, when the strategy underperforms, they conclude the system is flawed—without noticing that their own execution has already drifted away from the rules.
This is one reason a process journal is not a soft skill. It is a control system. If you record the signal, the rule, the market context, and the actual action taken, you can later separate strategy failure from execution failure. That distinction is often the difference between learning and self-sabotage.
Note
Many investors think they are being prudent when they stop a strategy after a drawdown. In practice, they are often reacting to pain, not updating a thesis.
What investors get wrong about abandoning a strategy
The biggest mistake is treating every bad stretch as a referendum on the strategy itself. A strategy can fail for at least four different reasons: the edge never existed, the edge existed but was overfit, the market regime changed, or the investor abandoned the process before the edge had time to express itself. Those are not the same problem.
This is where the literature on overfitting and survivorship bias becomes relevant. A backtest can look excellent because it fit noise, not signal, and a live strategy can look worse than it is if the investor only remembers the surviving winners [6][7]. If you want a deeper framework for this, AIBROKER’s guides on backtest checklist and survivorship bias are useful companions.
The honest assessment is that patience is not always a virtue. Some strategies deserve to be retired. The problem is that investors usually do not retire them using evidence. They retire them using discomfort. That is a very different standard.
Evidence checklist: continue, investigate, or stop
Check
Supports continued monitoring
Triggers investigation
Rule adherence
Executed as specified
Implementation drift
Point-in-time validation
Holdout behavior within plan
Live/holdout failure
Risk budget
Loss within precommitted bound
Bound breached
Market/data assumptions
Still valid
Structural or data change
Sample information
Threshold not crossed
Predefined threshold crossed
Decision matrix: bad streak or broken strategy?
This matrix is a process tool, not a performance guarantee. It is designed to separate emotional discomfort from evidence of structural failure.
A worked probability example, not a verdict on an edge
With independent binary trades and a constant 55% win rate, the loss probability is 45%. Five losses in one specified five-trade window have probability 0.45^5, about 1.85%. Across 100 trades, overlapping opportunities raise the exact probability of at least one such run to about 64.68%, calculated by a finite-state run recursion.
Worked independent-trial calculation
Input or result
Value
Meaning
Win probability
55%
Assumed constant
Loss probability
45%
Assumed constant
Five losses in fixed window
1.85%
0.45^5
At least one run in 100
64.68%
Overlapping run recursion
Expected return
Unknown
Payoff sizes missing
The calculation says nothing about expected return because win and loss magnitudes are absent. It also fails when trades overlap, regimes change, or outcomes are dependent. Use it to check intuition, then model the actual payoff sequence, costs, drawdown, and capital at risk.
DALBAR is a proprietary estimate, not causal proof
DALBAR reports compare modeled investor returns with benchmarks, but results depend on fund-flow data, timing assumptions, benchmark choice, and the period. A gap cannot by itself prove that the same investors bought high and sold low, and it should not be presented as a universal behavior penalty. Obtain the methodology for the cited edition before quoting a number.
The defensible lesson is narrower: cash-flow timing and investor decisions can make money-weighted experience differ from a benchmark’s time-weighted return. Measure your own deposits, withdrawals, fees, taxes, and policy deviations instead of borrowing an aggregate headline.
Discipline is not a personality trait. It is a system. The best investors reduce the number of decisions they have to make while emotional. They pre-commit, document, and define exit criteria before the drawdown arrives.
Pre-commitment devices work because they move the decision upstream. You decide in advance what counts as a normal drawdown, what counts as a review event, and what counts as a true failure. That way, the market does not get to negotiate with you in real time.
Tool
What it does
Best use case
Written max-drawdown rule
Defines when a strategy is reviewed
Systematic strategies with known risk bands
Cooling-off period
Delays impulsive changes
Discretionary investors prone to panic
Position-sizing cap
Limits damage while preserving exposure
Strategies with uncertain edge durability
Process journal
Separates rule adherence from outcome
Any strategy you may be tempted to abandon
Independent review trigger
Forces a second look before changing rules
When emotions are high and conviction is low
Checklist: pre-commitment devices for drawdown periods
A process journal is especially useful. Record the signal, the rule, the market context, and whether you followed the plan. Over time, you can distinguish between a bad strategy and a good strategy executed badly. That distinction is often the difference between learning and self-sabotage.
If you want a practical companion, AIBROKER’s dollar-cost averaging article shows how a rules-based contribution plan can reduce the temptation to time pain. The same logic applies to strategy management: consistency is not glamorous, but it is often the edge.
A precommitted review tree, not an instruction to hold
Separate implementation failure, risk-budget breach, invalid assumptions, and statistical underperformance. Define thresholds before deployment. Historical drawdown alone is not a safe boundary: backtests may be overfit and future losses can exceed every observation.
Most strategy abandonment happens in the ugly middle: after the honeymoon, before the recovery. That is when the strategy is no longer novel, but the evidence is still too thin to justify a verdict. A timeline helps investors avoid premature conclusions.
The timeline below is not a forecast. It is a behavioral map. It tells you what kind of response is appropriate at each stage of a drawdown so that you do not overreact to the first sign of pain.
Stage
Typical emotional reaction
Disciplined response
First few losses
Surprise
Check execution, not the thesis
Streak extends
Doubt
Compare results to historical drawdown bands
Drawdown becomes uncomfortable
Urgency
Reduce size only if risk budget is breached
Underperformance persists beyond review threshold
Concern
Investigate regime change, data issues, or edge decay
Evidence of structural failure
Relief mixed with regret
Retire or redesign the strategy
Timeline: how disciplined investors respond to a losing streak
What the research does not say
The honest caveat is that not every losing streak deserves patience. Some strategies are genuinely broken. Some are overfit. Some depend on market conditions that no longer exist. The research on loss aversion and the disposition effect does not tell you to hold forever. It tells you that your first impulse during pain is often unreliable.
That is why the right question is not, "Am I uncomfortable?" The right question is, "Has the evidence crossed the threshold I set before the discomfort started?" If the answer is yes, exit. If the answer is no, do not let a bad week become a bad decision.
This is also where a little humility helps. Investors often want a single rule that tells them exactly when to quit. Real markets do not offer that. What they offer is uncertainty, and the best response to uncertainty is a process that can survive being wrong for a while.
For investors who want to connect this to everyday portfolio behavior, AIBROKER’s article on dollar-cost averaging is a useful reminder that consistency often beats emotional timing, especially when markets are noisy.
So what
Losing streaks are not just a statistical event. They are a behavioral trap. The investor who survives them is usually not the one with the strongest opinions. It is the one who pre-decided what evidence matters, wrote it down, and refused to improvise under stress.
If you manage money systematically, your job is not to eliminate drawdowns. Your job is to make sure drawdowns do not force you into a worse strategy than the one you already had.
Closing perspective
Good strategies often die from bad timing, not bad logic. If you want to improve your odds, stop asking whether a drawdown feels bad enough to justify action. Ask whether your rules, your sample size, and your evidence justify it. That small shift—from emotion to process—is usually where discipline begins.
The practical move is simple: write the abandonment rule before the next losing streak starts, not during it. Then review it only when the evidence, not the adrenaline, says it is time.
Kahneman, D., & Tversky, A. (1979). Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2), 263–291.Source
Odean, T. (1998). Are Investors Reluctant to Realize Their Losses? The Journal of Finance, 53(5), 1775–1798.Source
Dalbar, Inc. (latest available). Quantitative Analysis of Investor Behavior (QAIB).Source
Feller, W. (1968). An Introduction to Probability Theory and Its Applications, Vol. 1 (3rd ed.). Wiley. Runs and streaks in repeated trials provide the probability basis for understanding clustered losses.Source
Grinstead, C. M., & Snell, J. L. (1997). Introduction to Probability. American Mathematical Society. Chapter on runs and repeated trials.
Barberis, N., Huang, M., & Santos, T. (2001). Prospect Theory and Asset Prices. The Quarterly Journal of Economics, 116(1), 1–53.Source
Statman, M. (1985). Disposition and regret in financial decision-making. Financial Analysts Journal, 41(6), 53–60.Source
U.S. Securities and Exchange Commission. Investor Bulletin: Understanding Behavioral Biases.