Survivorship Bias in Trading — The Error of the Survivors

It is late, you are scrolling, and the feed serves up the same thing as yesterday: a screenshot of an account up twenty thousand, someone bragging about a third green month in a row, comments underneath saying "I teach, DM me." Your own account is in the red after six months of learning, and a quiet, ugly thought arrives: "maybe I'm just the one who can't do this." You are not. You are seeing only the people who survived — and around them stretches a vast, invisible graveyard of accounts that blew up and went silent. This is survivorship bias, and it costs beginners more illusions than any other cognitive error.

What survivorship bias is, and where it comes from

Survivorship bias is the systematic habit of drawing conclusions only from the cases that "made it" to the moment of observation, while completely ignoring those that dropped out along the way. Because the casualties are invisible, the brain behaves as though they never existed — and builds its picture of the world from a sample made up entirely of winners.

The best illustration comes from the Second World War. The U.S. military analysed bombers returning from raids and mapped where the bullet holes landed — clusters on the wings, fuselage, and tail. The conclusion looked obvious: add armour exactly there. Statistician Abraham Wald, working in the Applied Mathematics Panel, reversed the reasoning. The analysis only ever saw the planes that came back. The holes in them marked the places a plane could be hit and still return. The areas to reinforce were the ones where returning planes were clean — the engines and the cockpit — because aircraft hit there simply did not come back to be counted. The same mechanism is at work when you learn trading exclusively from people who "made it home."

Where survivorship bias hides in trading

Social media is the loudest source. The algorithm pushes screenshots of wins because they trigger emotion and envy; a loss has no reach, so whoever lost usually deletes the account or goes quiet out of shame. You receive a stream made almost entirely of winners and naturally overestimate how easily and how often people win. Beneath that runs a second current: strategy stories. You hear "I made money on pattern X," but you never hear from the dozens who lost with the same pattern, because nobody advertises a loss.

Third comes mentor and guru advertising: the students it features are the ones it worked for, while those who dropped out after a few weeks never appear in the promotional material. Fourth — testimonials and screenshots on some broker sites, plus so-called verified accounts, which can be cherry-picked from the best or run on micro-balances for marketing. Fifth, and most insidious because it is technical: backtests. You optimise a strategy on the pairs and instruments that "survived" until today, and the historical data often omits currencies that were redenominated, hyperinflated, or vanished. That is a short road to curve-fitting your rules to the past of the survivors — and to a bitter surprise on the live market.

How it distorts your expectations and decisions

The effect is practical and expensive. First, you set an unrealistic target — if "everyone" is making tens of per cent a month, your own few per cent a year looks like failure, even when it is a respectable result. Second, you pick a strategy from an anecdote rather than from full-sample data that also includes the people for whom it did not work. Third, you trust anonymous, unaudited screenshots and copy other people's positions without knowing the full history of the account.

A natural trap of reasoning appears too: if someone turned a thousand into a hundred thousand, surely they had a method worth copying. Not necessarily. With enough people trying, plain statistics guarantees that a few will hit a spectacular run by pure chance — and it is precisely those few you will see at the top, because the rest went quiet. Daniel Kahneman calls this reflex "what you see is all there is": the mind builds a coherent story from the fragments available and never accounts for what is missing from the frame. It is one of the most dangerous psychological traps in trading, because it works quietly and feels like common sense.

The base rate — the number they never show you

The strongest antidote to survivorship bias is the base rate — the share of all participants to whom a given outcome happens, considered before you look at any single case. Here we actually know the number from a hard, regulatory source. When the European regulator ESMA introduced its restrictions on CFDs in 2018, it reported that typically around 74 to 89 per cent of retail client accounts lose money. That is why every broker in the European Union now displays a mandatory warning with its own specific percentage of losing accounts.

Imagine you land on a profile boasting "98 per cent win rate" and a run of nothing but green months. Set it against the base rate: most of retail loses. That clash does not prove the person is lying — it proves they are extremely atypical, and before you copy anything you must explain why they, specifically, would belong to the narrow minority. Most often it turns out you are looking at a short, cherry-picked slice or an ordinary lucky streak. The figures from individual accounts below are hypothetical, used only to show the mechanism.

A concrete defence against survivorship bias

The first rule is a verbal reflex: at every success story, ask "where is the graveyard?" Deliberately seek out those who dropped out — the "I blew my account" threads, the failure post-mortems, books about spectacular collapses such as "When Genius Failed" on the LTCM fund. The second: anchor your expectations to the base rate, not to the highlight reel. Since most people lose, your realistic early goal is not to double the account but to survive and learn slowly at small risk per trade.

The third rule is about data. Test a strategy out of sample — on a different period and different instruments, including ones that performed poorly or disappeared from the market. If the rules only work on a carefully chosen set of "surviving" pairs, you have fitted them to the past. The fourth: discount anonymous track records. A screenshot is not an audit; without a verified, multi-year account history, treat a great result as an unconfirmed anecdote, not as proof of a method.

"We see the winners and draw conclusions from their traits, never noticing the defeated — because the losers do not write memoirs. That silent majority, buried and invisible, we mistake for an absence of evidence that chance is at work." — Nassim Nicholas Taleb, Fooled by Randomness (Texere, 2001), paraphrasing the thesis on the hidden role of luck.

What to do this week — first steps

Instead of another evening of other people's screenshots, do three things. First: list ten profiles that recently inspired you, and against each note whether a verified, multi-year account history exists. Most of the sample will fail — and that is your lesson in proportions. Second: deliberately read five failure post-mortems from forums, or one book about a collapse, looking for recurring causes such as excessive leverage, no stop loss, or revenge trading. Third: write your goal for the coming year next to the ESMA base rate and check whether it is still realistic.

If a success story still tempts you to copy it, return to the question "where is the graveyard?" — and decide only once you can estimate how many people did the same thing and went silent. It is a simple change in the habit of looking, but it is the one that separates the trader who learns from full reality from the one who chases a slice made up of survivors alone.

Related reading: overconfidence bias — the twin error that makes you mistake a lucky streak for skill; herd mentality — why a stream of other people's wins pushes you into the crowd; realistic goals — how to anchor expectations to the base rate instead of the highlight reel; process over outcome — why a single spectacular result tells you nothing about the quality of a method.