Survivorship bias in trading: how hidden failures affect decisions

Survivorship bias can make trading results look more reliable than they are. It can appear in backtests, fund databases, index histories and trading education – especially when the failures are harder to see than the successes.

What is survivorship bias?

Survivorship bias happens when you judge a group by looking only at the examples that are still visible. In trading, that might mean focusing on companies that are still listed, funds that are still open, traders who are still active, or strategies that are still being promoted. The problem is what gets left out. Companies can be delisted. Funds can close. Traders can stop trading. Strategies can stop working. If these outcomes are missing from the data, the remaining examples can make success look more common than it really is.

A well-known example comes from the work of statistician Abraham Wald during the Second World War. The military wanted to reinforce returning bomber aircraft by looking at where the bullet holes were. Wald pointed out that these planes had survived, so the damaged areas were not necessarily the most vulnerable. The more important question was where the non-returning planes had been hit – but those planes were not available to inspect (JSTOR, accessed 12 June 2026).

The same idea applies in trading psychology. The results that are easiest to see are often the ones that survived. That does not make them false, but it can make them incomplete.

How traders develop ‘survivorship bias’

Survivorship bias is not just a mistake made by individual traders. It is also a feature of the information environment. Traders often see more success stories than failures because successful examples are easier to find, easier to promote and easier to remember.

The information landscape is filtered by survival

Much of the information available to traders comes from survivors. Trading books are usually written by traders who had enough success to attract an audience. Trading forums tend to highlight setups that worked recently. Fund databases may focus on funds that are still operating. Backtesting data may include companies that are still listed, while excluding those that were delisted or went bankrupt. None of this needs to involve deliberate distortion. The filter can happen naturally. But the result is still the same: successful outcomes become more visible than failed ones.

Success stories are easier to remember

Success stories are also more memorable. A trader who built a strong track record, a strategy that performed well, or a fund that beat its benchmark can make for a clear and compelling narrative. The full picture – including the many similar attempts that did not work – is usually less visible and less engaging. This can make successful outcomes feel more common than they are. A trader may remember the success stories they have read, while the missing failures remain outside their field of view.

Failure data is harder to find

Failure data is often less accessible. Traders who lose money, abandon strategies or leave the market rarely publish detailed accounts of what happened. Failed funds may disappear from databases. Delisted companies may be harder to include in standard backtests.

Types of survivorship bias in trading

Survivorship bias can appear in several trading contexts. The same basic problem applies in each case: the visible sample does not include everything that should be counted.

Past performance is not a reliable indicator of future results.

Survivorship bias in practice: trading examples

The easiest way to understand survivorship bias is to see how it changes the interpretation of real trading information.

The backtested strategy that seems to work

A trader tests a momentum strategy on a dataset of stocks covering the last 20 years. The results look strong and consistent. But if the dataset includes only companies that survived the full period, it may exclude stocks that went bankrupt, were delisted, or failed during those 20 years. These companies would have been part of the actual trading universe at the time, so leaving them out can make the strategy look more reliable than it was. The opposite can also be true: not every missing company performed poorly. Some may have been delisted after mergers or acquisitions. That’s why a survivorship-adjusted dataset, including failed and delisted companies where possible, can give a fuller view.

The successful trader whose method appears reliable

A trader reads a book by someone who turned a modest starting amount into a much larger sum over several years. The method is clearly explained, and the results are real. What’s missing is the wider group of traders who tried a similar method over the same period. Some may have succeeded. Others may have lost money, changed strategy, or stopped trading. This doesn’t make the story misleading on its own. The opposite may also be true: the trader’s method may have been disciplined, repeatable, and well suited to the market conditions at the time. But it should still be treated as one example, not proof that the same approach is likely to work in the same way for others.

The hedge fund with impressive historical returns

A trader reviews a hedge fund database and sees that the average fund produced returns above its benchmark. But if the database includes only funds that are still operating, the result may be incomplete. Funds that closed due to poor performance may be missing. If they were included, the average return might be lower. This gap between the visible result and the fuller result is the effect of survivorship bias. The opposite scenario is also possible: some funds may close for reasons unrelated to performance, such as mergers, strategy changes, or investor withdrawals. So survivorship bias doesn’t automatically invalidate the data. It means the trader should check what is included, what is excluded, and how that might affect the conclusion.

How survivorship bias affects your decisions

Survivorship bias matters because it can influence the assumptions traders use when making decisions.

• Overestimating probability: seeing mostly successful examples can make a strategy look more reliable than it is. This can affect position sizing, risk tolerance and expectations.
• Using misleading benchmarks: comparing your results with visible winners, fund averages or trading community results may mean comparing yourself with a filtered group.
• Choosing strategies based on success stories: a timeframe, indicator, market or risk setting can look attractive when it appears in strong examples, but those examples may not show how often similar attempts failed.

Survivorship bias risks in leveraged trading

Survivorship bias is especially important in CFD trading because leverage can amplify both gains and losses. If a trader overestimates the probability of success, they may take on more exposure than the strategy can realistically support.

A backtest may show a strategy returning 15% a year, for example, but if that result came from survivor-only data, live results may be materially different. The reverse can also happen: if a backtest focuses too heavily on failed examples or difficult market periods, a strategy may look less robust than it really is. In both cases, filtered data can distort expectations.

When leverage is involved, the gap between the backtested result and the live outcome can have a larger effect on account equity.

How to correct for survivorship bias in trading

Correcting for survivorship bias means asking a simple question: what is missing from the data?

• Step 1: Use datasets that include delisted companiesWhen backtesting equity strategies, use data that includes delisted, bankrupt and acquired companies, not just companies that are still listed. This gives a fuller view of what the trading universe looked like at the time. Survivorship-adjusted datasets can be more complex to work with, but they help reduce the risk of building a strategy around cleaner historical data than a trader would have had in live markets.
• Step 2: Look for failure data, not only success dataWhen assessing a strategy, fund or trader, look for the outcomes that did not work as well as the ones that did. If a strategy is said to have worked in 70% of historical cases, check how the full sample was defined and what happened in the other 30%. This does not mean ignoring successful examples. It means putting them in context.
• Step 3: Read the regulatory profitability disclosures
  CFD brokers regulated in certain jurisdictions, including the EU and the UK, are required to disclose the proportion of retail clients who lose money trading CFDs on their platform. These disclosures can help traders set more realistic expectations. They show that trading outcomes vary widely, and that losses are common among retail CFD traders.
• Step 4: Treat success stories as examples, not proof
  Successful traders, strategies and funds can offer useful lessons. But each success story is still only one example. When reading about a strong result, ask what wider group it came from. How many traders attempted something similar? How many funds used a similar approach? How many strategies were tested but not published? These questions help place the visible success in a fuller context.
• Step 5: Use realistic base rates
  A base rate is the broad success or loss rate for a group before looking at a specific example. In trading, base rates can help keep expectations grounded. For example, if research or regulatory data shows that many retail traders lose money, that should inform how a trader thinks about risk, even if they have read several strong success stories. Base rates do not predict any single trader’s result, but they can help avoid overly optimistic assumptions.
• Step 6: Keep a full trading journal
  A trading journal can help traders build their own evidence base. The most useful journals track all setups, not only the trades that worked. This can include trades taken, trades skipped, reasons for entry, reasons for exit, and whether the setup met the original plan. Over time, this creates a clearer picture of how a strategy performs across both positive and negative outcomes.

Developing psychological awareness can support more disciplined decision-making, but it does not remove the risks inherent in CFD trading.

Common mistakes when addressing survivorship bias

Survivorship bias is relatively straightforward to understand, but harder to avoid in practice:

• Assuming visible data tells the whole story: stock databases, fund rankings and trading education may show what survived, not what failed, closed or disappeared. Ask what had to happen for an example to become visible.
• Dismissing all historical data: past data can still be useful. The aim is to understand its limits, not ignore it completely.
• Knowing the bias but not changing the process: awareness only helps if it affects how you test, compare and interpret results. That may mean using fuller datasets or treating published results with more caution.
• Comparing yourself with visible peers: trading communities and competitions may overrepresent traders who stayed active long enough to be seen. Peer comparison can still help, but it may not reflect the full group that started out.

Survivorship bias can’t always be removed completely. It’s one part of realistic risk assessment, and works best when considered alongside fuller data, consistent process discipline and balanced expectations.

Survivorship bias and risk management

Survivorship bias can affect risk management because it can inflate the assumptions behind position sizing, stop-loss placement and return expectations.

Create an account Open a demo account

FAQ