Psychology

Survivorship Bias: Why It Feels Like Everyone Else Is Winning

Survivorship bias is what happens when you judge your results against a sample that's secretly made of only the winners — dead coins delisted, blown-up traders gone quiet, failed funds dropped from the record. Why your feed makes your losses feel uniquely bad, and how to check the real numbers before you believe it.

By Vijay Rathod

Survivorship bias is the error of judging your own results — or an asset, a fund, or a strategy — against a sample that quietly excludes everything that failed. In investing, it shows up as an asymmetry: the losers get delisted, deleted, and abandoned into silence, while the winners stay visible in your feed, in “top gainers” lists, and in historical charts. The comparison you’re making in your head was never fair, because half the sample was removed before you ever saw it.

This matters most right after a loss, when it feels like proof of personal failure that “everyone else” seems to be up. They’re not. You’re looking at a graveyard with the headstones removed.

Where the missing losers actually go

In crypto, they go dead. CoinGecko’s research team tracked cryptocurrency projects through the end of 2025 and found that 53.2% of all tokens ever launched — roughly 11.6 million out of about 20.2 million projects — are now classified as “dead coins”: abandoned, valueless, or with trading volume that’s effectively zero. 2024 alone produced roughly 1.4 million of those failures, driven partly by how cheap it became to launch a token on platforms like pump.fun. Every one of those dead projects had a Twitter account, a Discord, and early holders posting screenshots of gains on the way up. Almost none of them are still posting on the way down. They just went quiet, and quiet doesn’t show up in a feed the way a green candle does.

In traditional markets, the losers don’t vanish from existence — they vanish from the historical record used to judge performance. This is the original, most rigorously studied version of the bias: mutual fund databases used to routinely drop funds after they closed or merged away, which meant only the survivors were left to calculate the industry’s “track record.” Economists Edwin Elton, Martin Gruber, and Christopher Blake measured this directly and found it overstated the U.S. mutual fund industry’s reported returns by about 0.9 percentage points a year — enough, on its own, to make an industry that actually underperformed its benchmark after fees look like it was roughly keeping pace. A related effect shows up in stock indexes themselves: a backtest run only on companies still in the S&P 500 or Nasdaq-100 today silently excludes every company that got delisted for going bankrupt or being acquired out of distress, which is exactly the kind of company most likely to have been a bad trade.

Either way, the mechanism is identical: the sample you’re checking yourself against has already had its worst outcomes edited out.

Why your feed makes this so much worse than a stale database

A dropped fund from a 1990s database is a passive, structural bias. Social media adds an active one on top of it, because posting is a choice, and the choice isn’t random.

  • People who are up post screenshots. People who are down mostly don’t — or they delete the account entirely.
  • Winners get amplified by engagement, so a single well-timed trade reaches you dozens of times through retweets and reposts, while a hundred quiet losses reach you zero times.
  • “Top gainers” and “biggest movers” lists are built, by design, to show you the extreme right tail — never the extreme left tail of things that went to zero.

None of this requires anyone to lie. It only requires that people who lost money go quiet more often than people who made money, which they reliably do. The result is a feed that is a biased sample of outcomes, dressed up as a representative one.

What this actually costs you

The direct cost isn’t the bad feeling — it’s the decisions the bad feeling causes. If your baseline belief is “everyone else is getting rich off this and I’m the one screwing it up,” the two most common responses are both expensive: chase a position sized way beyond your plan to catch up to a “normal” gain that was never actually normal, or conclude you’re uniquely bad at this and quit at exactly the point where a calmer, smaller version of the same strategy might have worked. Related patterns covered elsewhere on this site make more sense once you see the biased sample underneath them — FOMO is what it feels like to chase the visible survivors, and herd mentality is what it feels like to follow them once enough other people are chasing too.

A worked example

Say you bought into a new token at launch alongside four friends, and it’s now down 80%. If you only checked Crypto Twitter, you’d see a wall of people up big on similarly “early” plays and conclude your specific pick was uniquely bad, or that you’re bad at this. But CoinGecko’s data says over half of everything launched in the same window as your token failed outright — not underperformed, failed completely. Your actual reference class isn’t “people posting gains today.” It’s “everyone who took a similar-stage bet,” and in that reference class, a loss is the modal outcome, not an outlier. That doesn’t make the loss feel good, but it changes the question from “what’s wrong with me” to “was the position sized like a bet with a real chance of failing” — which is a question you can actually act on.

Three checks before you let a comparison change your behavior

  1. Ask what you’re not seeing. For every winner in your feed, some number of similar bets failed silently. You don’t need the exact number to remember the sample is incomplete by construction.
  2. Find the base rate, not the anecdote. A published failure-rate study (like the CoinGecko figures above) or a fund’s actual long-run return after outflows tells you more than any number of individual screenshots, because it includes the losers.
  3. Separate “this feels true” from “this is representative.” A feed built on selective posting will always feel true in the moment — that’s what makes it convincing. Representative is a different, checkable claim, and it’s usually the one that’s false.

FAQ

Isn’t it just normal to compare myself to other investors? Comparison itself isn’t the problem — the sample you’re comparing against is. A fair comparison uses everyone who took a similar risk, winners and losers together. Your feed, your group chat, and even historical index charts systematically drop the losers, so you’re not comparing yourself to “other investors” — you’re comparing yourself to a curated subset of the ones who happened to survive.

How is this different from FOMO or herd mentality? FOMO and herd mentality are about what you do in the moment — chasing a price or following a crowd. Survivorship bias is about what data reaches you in the first place. It’s upstream of both: a feed made of only winners is exactly what manufactures the fear of missing out and the sense that “everyone” is doing something you’re not.

Does this mean the assets or strategies that survived aren’t actually good? Not necessarily — some survivors genuinely earned it. The point isn’t that nothing that lived is worth anything; it’s that a pile of survivors alone can’t tell you the odds you were actually facing, because you never see the pile of everything that started out looking just as promising and didn’t make it. Judge a strategy by its odds going in, not by pointing at the few names that came out the other side.

If the gap between how you feel and what actually happened is a recurring problem, not a one-off, Loss Aversion Is Making You Worse at This covers the broader wiring behind why losses distort judgment more than an equivalent set of facts should.