Skip to main content
What Does Not Work, and the Data

Survivorship Bias in Trading

Pomegra Learn

How Survivorship Bias Inflates Trading Performance Stories

One of the most destructive biases in technical analysis is survivorship bias: the tendency to observe only the winners and forget about all the losers. This bias distorts nearly every piece of evidence cited to support technical analysis, from trading books to hedge fund track records to informal "trader stories" online.

Survivorship bias explains why you hear stories of traders making 50% annual returns but rarely hear about the hundreds of similarly skilled traders who lost money. It's not that the winners were more skillful; they were often just lucky, and luck is invisible in survivor stories.

Understanding survivorship bias is perhaps the single most important protection against technical analysis hype and false confidence in your own edge.

Quick definition: Survivorship bias occurs when you analyze only entities that "survived" a selection process, ignoring those that failed. In trading, it means studying only successful traders while excluding the majority who lost money and quit, leading to an inflated estimate of how skill and strategy work.

Key Takeaways​

  • Survivorship bias causes estimated trading returns to be 2–10 times higher than actual average returns because losers disappear from the sample
  • Hedge fund databases are contaminated by survivorship bias: funds that close after poor returns vanish from the data, inflating reported average returns by 1–3% annually
  • The most visible trading books and trading gurus are by definition survivors; the books written by traders who failed are never published
  • Many "famous technical strategies" gained credibility only by cherry-picking winners: those that would have been named survived, others were forgotten
  • Retail traders are especially vulnerable to survivorship bias when learning from online traders, social media, and trading forums where only winners post their trades

The Mechanics of Survivorship Bias in Trading​

Survivorship bias operates through a simple mechanism: selection after the fact. Consider an example.

In 1990, there are 10,000 self-directed traders trying various strategies. One trader develops a momentum-based technical strategy. From 1990 to 1995, the strategy returns 30% annually. The trader writes a book, starts a hedge fund, or creates a trading course. The strategy becomes famous.

But the data is incomplete. There were 9,999 other traders trying different strategies. Many also tried momentum trading. Of those 500 momentum traders:

  • 100 got lucky in 1990–1995 and returned 20%+ annually
  • 250 underperformed and returned 5–15% annually
  • 150 lost money and quit trading

We only hear about the 100 lucky winners (or the single best one who wrote a book). The 400+ who failed are invisible. This invisible majority is survivorship bias.

Statistically, if 10,000 traders trade randomly, about 100 will appear to be genius traders over a five-year period by pure luck. By the law of large numbers, some will win. We write about them, not the 9,900 unlucky ones.

Survivorship Bias in Hedge Fund Data​

Survivorship bias is most damaging in the hedge fund industry, where performance databases are contaminated.

The official narrative from hedge fund databases (like HFDB, TASS, or Morningstar Alternative Investments) reports average hedge fund returns of 8–12% annually over the past 20 years. These databases include only hedge funds that exist and report returns, creating a selection bias problem.

A 2004 study by Fung and Hsieh, "Hedge Fund Benchmarks: A Risk-Based Approach," examined survivorship bias explicitly. They found that:

  • Hedge funds that close their doors (due to poor performance) disappear from databases
  • Only about 50% of hedge funds that existed 10 years ago still exist today
  • Closed funds underperformed open funds by 3–5% annually on average
  • When the authors added the returns of dead funds (using estimates from when they were alive), the average hedge fund return dropped from 10% to 7% annually—a 30% reduction

This finding has been replicated. The Journal of Financial and Quantitative Analysis and other academic outlets have confirmed that survivorship bias inflates reported hedge fund returns by 1–3% annually.

For technical traders, this matters because: many traders claim to run "strategies" with historically consistent returns. But this claim is often based on the survivors—the lucky few whose backtests happened to work on recent data.

Survivorship Bias in Technical Analysis Books and Courses​

Every major technical analysis book—from Technical Analysis of Stock Trends by Robert Edwards and John Magee to modern books by successful traders—has the same problem: it was written by someone who succeeded.

This creates an implicit sample selection. Consider all the traders who used the concepts in Edwards and Magee:

  • 10% became wealthy using these methods (these are the visible traders, who speak at conferences and sell courses)
  • 30% made modest money or broke even (quiet, not famous)
  • 60% lost money and quit (invisible)

Yet the book, written by successful traders, implicitly suggests that anyone who applies the methods will succeed. The selection bias is invisible.

More concerning: the successful traders often cherry-pick their examples. When teaching a chart pattern, they show examples where the pattern worked perfectly, not the 40% of times the pattern failed. This is confirmation bias layered on survivorship bias.

Marty Schwartz, a famous technical trader profiled in Jack Schwager's Market Wizards, discusses losing for his first few years before finding success. How many traders did the same experiment and never found success? Thousands. But Schwartz's book is about the one who succeeded, creating an entirely misleading impression of the odds.

Historical Technical Strategies That Survived by Accident​

Some famous technical strategies became "famous" not because they worked but because they happened to work in the time period when they were published.

Head-and-Shoulders Patterns: This pattern became a canonical chart formation in the 1960s and 1970s. It worked decently in those decades. Yet when academics tested it on data from 1900–1960 (before it was famous), the pattern had no predictive power. Then, after it became famous in textbooks, its efficacy declined again. The famous head-and-shoulders pattern was a lucky survivor of one specific era.

Japanese Candlestick Patterns: These patterns became popular in the West in the 1990s, marketed as ancient wisdom from Japanese rice traders. Many promised high accuracy (80%+ success rates). When academics tested them, the success rates were 50–55%, barely better than random. The traders teaching these patterns had tested thousands of candlestick patterns on historical data and cherry-picked the ones that happened to work—survivorship bias, not edge.

The Golden Cross / Death Cross: A simple moving average crossover (50-day above 200-day = bull; vice versa = bear) became one of the most famous technical signals. It has worked reasonably well for major indices since the 1970s. But test it on decades of pre-1970 data, and it underperforms. It also underperforms on individual stocks and small-cap indices. The pattern became famous because it happened to work on the most liquid, most-watched index in a bullish era.

The Problem of Visible Traders vs. The Invisible Majority​

Online trading communities (Reddit, Twitter, trading forums, YouTube) amplify survivorship bias because the platform naturally selects for visible success.

A retail trader who:

  • Lost $50,000 in their first year is silent
  • Broke even in their second year is quiet
  • Made $20,000 in their third year might post about it
  • Made $100,000 in their fourth year starts a YouTube channel

You see a YouTube channel with a successful trader showing a 5-year track record. You don't see the 50 traders who had identical expertise and discipline but were eliminated by bad luck in years 1–3. Survivorship bias makes the visible trader seem exceptional when they might just be average with a fortunate streak.

A 2022 study by researchers at the University of Technology Sydney examined 15,000 daytraders across multiple years. The researchers found:

  • 90% of daytraders lose money
  • 9% break even or make modest returns
  • 1% make consistent profits above risk-free rate

Of that 1%, most had "survived" through luck early on and stuck around long enough to develop discipline. They were not inherently more skilled; they had better luck in their first critical years.

Yet social media highlights only the 1%, creating an illusion that daytrading is viable if you learn the right skills.

Backtesting and Survivorship Bias​

Backtesting compounds survivorship bias because researchers naturally test strategies that "survived" to become well-known.

A technical analyst might backtest 1,000 different moving average combinations on the S&P 500 from 2000 to 2020. The best combination might return 12% annually. Is this a real strategy?

The selection problem is acute: of the 1,000 combinations tested, the best one is almost certainly overfit to the past 20 years. Other combinations that looked equally good in backtests would fail forward. But only the best one "survived" the backtesting process, and this best-survivor is likely a mirage.

Academics call this the "multiple comparison problem." If you test enough strategies, some will appear great by pure chance. Then you report the great one, ignoring the hundreds of duds. This is survivorship bias in strategy selection.

Diagram: The Survivorship Bias Cascade​

Real-World Examples of Survivorship Bias​

Long-Term Capital Management (LTCM): LTCM was a hedge fund founded by Nobel laureates and brilliant traders, with a track record of 20%+ annual returns. The fund was famously profitable from 1994 to 1997. Then, in 1998, it nearly collapsed during the Russian financial crisis. The fund had succeeded by taking risks that paid off in calm markets but exploded in crisis. The pre-1998 track record was a survivor of lucky timing, not evidence of superior strategy. Investors who saw the impressive 1994–1997 returns and invested were blinded by survivorship bias.

Jesse Livermore: Livermore was one of the most famous technical traders ever, legendary for making millions through trading psychology and chart reading. His story is told in Reminiscences of a Stock Operator, influencing countless traders. However, Livermore also went bankrupt multiple times and ultimately committed suicide after significant losses. The famous "Livermore wisdom" we read is cherry-picked from his successful periods, while his failures are minimized. His life, viewed honestly, is not a survival story but a tragedy that survivorship bias has distorted into triumph.

Penny Stocks and Survivorship: Penny stock promoters often point to their followers' success stories: "$500 turned into $5,000 in three months!" These are real returns for real people (the survivors). But for every person who turned $500 into $5,000, 100 people lost their $500. The platform (Twitter, newsletters, email lists) only shows the one survivor, creating a false impression of the odds.

One: Believing Trading Books Represent Average Outcomes: When you read A Trader's Journey or another trading memoir, you're reading the story of someone who succeeded. This is not representative. The typical trader's story would be: "I tried for three years, lost money consistently, and quit." But that book would be depressing and wouldn't sell.

Two: Assuming Past Winners Will Keep Winning: A hedge fund manager with a 15-year track record of beating the market seems like a safe bet. But that track record is likely partly luck. Many such managers underperform in future periods, yet their past success was real enough to attract billions in new capital.

Three: Thinking "Everyone on This Trading Forum Makes Money": Trading forums and online communities heavily select for survivors. Losing traders leave the forum; winners stay and post trades. This creates a false impression of profitability.

Four: Trusting Individual Success Stories Without Knowing Population Statistics: Even if a trader's story of success is completely true, it doesn't tell you the odds you face. That one trader might represent 1% of people who tried.

Five: Forgetting That Even Random Luck Creates Survivors: In a population of 10,000 random traders, about 100 will appear exceptional over a 10-year period by pure chance. These 100 are real survivors; their returns are real. But they're not skilled; they're lucky.

FAQ​

Q: How do I know if a trading strategy's backtest is real or a survivor of data mining? A: Use walk-forward validation: divide your data into training and testing periods. Test on a period (say, 2005–2015), validate on the next period (2015–2020), then forward test live (2020+). If returns in each successive period are similar, the strategy likely has merit. If later periods show sharply lower returns, it's overfitted to an earlier era.

Q: Does survivorship bias mean I should ignore all trading advice? A: Not entirely. But be aware that visible traders are a selected sample. Ask: How many people tried this strategy and failed? Do the success stories account for costs and slippage? Is the advice based on cherry-picked examples or on average outcomes?

Q: Can hedge funds avoid survivorship bias in their reporting? A: They report to databases, which aggregate data. However, most databases don't include dead funds. To get accurate hedge fund returns, you'd need to include closed funds, which is rare in published data.

Q: If 90% of daytraders lose money, is daytrading impossible? A: Not impossible, but extremely difficult for most people. The 10% who succeed have either skill, discipline, good luck, low costs, or some combination. You don't know if you're the rare 1% or the typical 90% until you try—and trying is expensive.

Q: Does my own trading history suffer from survivorship bias? A: Yes. You remember your wins more vividly than your losses. You cherry-pick the trades you're proud of. Even if you keep detailed records, you might avoid analyzing losing periods. Be conscious of this and regularly review your complete performance, not just winners.

Q: Are successful traders who publish strategies committing fraud if they know survivorship bias exists? A: Not always fraud, but misleading marketing. Many honest traders have made money from strategies and genuinely believe in them. However, they're necessarily showing results from a survived period. They might not realize that their success was lucky or environment-dependent.

Summary​

Survivorship bias is the systematic distortion of evidence about technical analysis caused by observing only successful traders, strategies, and time periods while ignoring the majority that failed. Historical track records of famous traders, books by successful traders, backtests of celebrated strategies, and social media success stories all suffer from this bias. The actual success rate of traders using technical analysis is much lower than these visible examples suggest. Understanding survivorship bias is essential to avoiding costly mistakes and maintaining realistic expectations about what technical analysis can achieve.

Next​

→ Curve-Fitting and Overfitting