Skill Ratings Inflate 22% After a 3-Win Streak—Then Losses Follow
The claim is arresting: self-assessed skill climbs roughly a fifth after just three consecutive wins, and the subsequent losses arrive with almost mechanical regularity. Anyone who trades altcoins will recognise the shape of that curve, even if they have never placed a wager in their life. The question worth asking is not whether the effect is real, but why a short run of success so reliably corrupts the judgement of otherwise careful people.
The Psychology of a Streak
Three wins is a trivially small sample. Yet the human brain treats it as signal, not noise. Daniel Kahneman and Amos Tversky's work on the law of small numbers showed that people routinely infer stable properties from tiny, unrepresentative samples. A handful of outcomes becomes a story about underlying ability.
This is compounded by variable-ratio reinforcement, the schedule famously identified by B.F. Skinner in which unpredictable rewards produce the most persistent behaviour. Crypto markets deliver exactly this pattern: irregular, high-salience payoffs that feel earned even when they are largely luck.
The 22% figure will not hold across every population, but the direction is robust. Confidence inflates faster than competence, and the gap has to close eventually.
Where the Inflation Bites in Altcoin Trading
Altcoin markets are unusually efficient at manufacturing streaks. Low liquidity, high volatility and round-the-clock trading mean a retail participant can string together three profitable trades in a week and conclude they have an edge.
Consider a realistic case. A trader buys three mid-cap tokens during a broad market rally, each up 15–40% within days. They begin sizing up, skip the due diligence they applied to the first position, and start treating a research checklist as bureaucracy. When the rally reverses — as it did sharply in 2022 and again in several 2024 altcoin drawdowns — the same positions fall further and faster than the majors. The inflated rating meets the correction.
There is a structural reason for this. Altcoins have thinner order books and higher beta to Bitcoin, so gains and losses both arrive amplified. A streak built on market beta is easily mistaken for selection skill.
Loss Aversion and the Refusal to Downgrade
The damage is not confined to overconfidence. Kahneman and Tversky's prospect theory describes loss aversion: losses hurt roughly twice as much as equivalent gains please. After a streak, this creates a specific trap.
A trader who has privately rated themselves highly cannot easily accept that the rating was inflated. Admitting error means accepting a loss on both the position and the self-image. The result is the disposition effect — holding losers too long, cutting winners too early — documented extensively in behavioural finance.
The 22% inflation is therefore not merely a mood. It is a decision input, and a corrupted one.
Designing Against Your Own Streaks
The practical response is procedural rather than motivational. Three interventions follow directly from the research.
Pre-commit to position sizing rules before a streak begins, and treat any deviation as a red flag rather than a signal. If your standard allocation is 2% of portfolio and a winning run tempts you to 6%, the temptation itself is the data point.
Keep a written decision log. Record the thesis, the evidence and the expected outcome for each trade. Reviewing it after a losing sequence is far more instructive than reviewing it after a winning one, because it separates process from outcome.
Use base rates deliberately. Ask what proportion of your last fifty trades were profitable, not your last three. Small-sample intuition is the enemy here, and a simple spreadsheet is the antidote.
What to Watch Next
As altcoin markets mature and retail access widens through regulated exchange services, the behavioural edge will matter more, not less. Products become easier to trade; discipline does not become easier to acquire. The traders who survive the next cycle will be those who treat a three-win streak as a prompt to audit their process rather than a licence to enlarge it. Track your rating honestly, and expect it to be wrong precisely when it feels most justified.