A Good Track Record Is Not Evidence of Skill
핵심 요점
- Survivors are manufactured by the sample size. Ten thousand coin-flippers guarantee a few perfect five-year records, and those people will not feel lucky.
- The path that happened is one draw from many that could have. Judging a decision by which draw appeared is judging the wrong thing.
- Checking more often makes you sadder without making you better informed — at short intervals you are almost entirely sampling noise.
학습 경로Judge the decision, not the outcome13단계 중 2단계
이 글보다 먼저:Resulting: Why a Winning Trade Is Not Proof of a Good Decision
참고 도서 Fooled by Randomness — Nassim Nicholas Taleb, 2001
Where star performers come from
You do not need any skill in the system to produce people with outstanding records. You only need enough people.
The point is not that every good record is luck. It is that a good record, on its own, cannot distinguish between the two — and the industry only ever shows you the survivors. The funds that closed are not in the advertisement, and the traders who blew up are not writing threads about their process.
This is why the honest question about any track record is not "how good is it?" but "how many people were running something similar, and where are they now?" A 1-in-500 record drawn from a pool of 5,000 attempts is the expected outcome, not an achievement.
The history that happened, and the ones that did not
A decision should be graded on the range of outcomes it exposed you to, not on which one arrived.
Someone who puts their savings into a single stock and triples it has a good outcome from a decision that also contained ruin. The tripling is the only branch anyone sees, including them — which is how the habit gets reinforced rather than corrected.
Why checking your portfolio hurts
Taleb makes an argument here that is unusually concrete for the book, and it changes behaviour immediately once you see it.
Because losses are felt more sharply than equivalent gains, a portfolio that drifts upward over a year delivers a stream of small painful moments to someone watching it hourly, and a single pleasant one to someone checking in December. Same portfolio, same return, opposite experience.
Checking constantly
- Mostly noise, felt as information
- Many small losses experienced
- Constant pressure to act
Checking rarely
- Higher share of real signal
- Fewer, larger, clearer moves
- Decisions made on a schedule
The record that looks safest
A long run of small steady gains is not evidence of low risk. Sometimes it is the risk.
The uncomfortable part is that this profile scores well on almost every measure someone would use to evaluate it. Low volatility, high consistency, an excellent Sharpe ratio, forty-eight winning months. All true, all measured on the period before the thing it was exposed to happened.
You cannot fully solve this, and Taleb does not claim to. The usable habit is to ask of any smooth return: what event is this strategy short? If you cannot name it, you have not established that there isn't one.
The part of this book to leave behind
The book's tone invites a conclusion it does not actually support: that since everything is luck, planning is pointless and everyone successful is a fraud. That reading is comfortable because it excuses you from doing the work.
The argument is narrower and more useful. Randomness dominates over short horizons and small samples; it does not dominate everything. Costs are not random. Position size is not random. Whether you diversify is not random. The correct response is to spend your effort on the parts that are not luck, which is also what the index argument rests on.
이번 주에 해볼 것
- Take a fund or trader whose record impressed you. Find out how many similar funds launched in the same year and how many still exist.
- Count how many times you checked your portfolio last week, and what you did differently as a result.
- Write down your best investment decision of the year and list three ways it could plausibly have gone badly.
- For any smooth-looking strategy you hold, name the event it would lose badly on.
자주 묻는 질문
What is the main idea of Fooled by Randomness?
That people systematically mistake luck for skill, because they see only the outcome that occurred and only the participants who survived. A track record on its own cannot separate a good process from a fortunate sequence.
What is survivorship bias in investing?
Judging performance from the funds, stocks or traders still visible while the failures have quietly disappeared. It makes the average look far better than the experience of everyone who actually started.
How long does a track record need to be to prove skill?
Longer than most people assume, and it depends on how variable the returns are. For typical equity strategies, distinguishing genuine skill from luck with confidence can take decades of data, which is why shorter records prove much less than they appear to.
Why does checking my portfolio less often improve returns?
Short intervals are dominated by noise, and because losses are felt more strongly than gains, frequent checking produces a stream of unpleasant impressions that push you to act. Fewer observations means fewer unnecessary decisions.
Does this mean investing is all luck?
No. Randomness dominates short horizons and small samples, but costs, diversification, position size and how long you hold are all under your control. The lesson is to put effort where luck is not the deciding factor.