Short answer: recency bias in trading is the habit of letting the most recent trades, candles or sessions count for more than the full record justifies. Five losers in a row feel like a broken strategy, and a month of trending price feels like a permanent regime. The fix is mechanical rather than mental: set your lookback window in advance, size from the long sample, and change rules only on a schedule. The readers in mind are CFD traders whose rules bend after every streak, on forex, indices, metals or crypto.
What recency bias is
Recency bias is one of the cognitive biases studied in cognitive psychology, and it is among the most ordinary. The mind treats fresh information as more representative than older information. In markets, the last few trades or the last few weeks of market trends end up shaping a decision that should rest on hundreds of observations.
Recency is a problem of weighting, not of memory. Our guide to trading psychology treats it as a state to monitor rather than a flaw to be ashamed of. A plain test: if your plan for the next session would look different had the last three trades gone the other way, recency is already in the chair.
How the mind weights fresh information
Three mechanisms do most of the work, and knowing them makes the bias easier to catch.
- Availability. A vivid, recent loss is quick to recall, so it feels more probable than it is. A stale loss from four months ago has faded, even if it was larger.
- Small-sample reading. Three or four outcomes get read as a trend when they are scarcely distinguishable from noise. A coin lands four heads in a row roughly one time in sixteen.
- Adaptive forecasting. The mind quietly extends the latest conditions forward. After a sluggish week it expects a sluggish day; after a violently trending week it expects more trend.
Price is unforgiving because it rarely behaves like weather. One losing day is one datum, not a verdict on the setup.
Signs it is steering your trading decisions
Recency bias seldom announces itself. It arrives as small, reasonable-sounding adjustments to your trading decisions:
- Position size drifts up after two or three winners and down after two or three losers, with no change in the written plan.
- You drop a setup after a short losing cluster, then take it again a month later when it has "started working".
- Five green 15-minute candles in a row trigger a market order at the high, with no setup behind it.
Each of these has a cost beyond the result. Chasing the latest mover means paying the spread and commission charge on trades the plan never called for.
Where it bites hardest: a market-by-market view
The bias is universal, but its favourite trigger differs by market.
| Market | Typical recency trap | What the longer record shows |
|---|---|---|
| Forex majors | Treating a two-week trend as a new regime and dropping range setups | Ranges return; lopsided trend months are a minority |
| Stock indices | Buying the fifth green day because "it only goes up" | Streaks are ordinary; entering at the high adds no edge |
| Gold and silver | Sizing up after one violent news candle paid off | Long wobbly, sideways stretches between those candles |
| Crypto (BTC, ETH, SOL) | Switching from swing to scalp after a jittery weekend | Volatility is lumpy; the next week is often unremarkable |
Index streaks are the most seductive. Eight green days feel like information; across the long record they are an ordinary event that says little about day nine.
A worked example: forty trades and two size swings
Suppose a setup has produced plus 0.25R per trade across 120 trades, and the trader risks 200 per trade, 1 percent of a 20,000 account. The next six trades deliver five losers and one 2R winner, a net of minus 600. The setup has not changed; the sample has delivered a normal bad cluster.
The recency-driven response is to halve risk to 100 "until it recovers". The next twenty trades produce plus 8R, so 100 risk earns 800 instead of 1,600. Confidence returns, risk goes to 400 "because it is working", and the following twenty trades deliver minus 4R. At 400 risk that costs 1,600 instead of 800.
At fixed size those forty trades would have made plus 4R, or 800. The recency-driven version lost 800. Two size swings, each triggered by the last handful of results, squandered 1,600 between the two paths, and not one bad setup was taken.
Common mistakes recency bias produces
- Retiring a strategy on fewer than 30 trades of evidence. Trading strategies tested across a long sample deserve the same sample before they are dropped. Thirty is a rough floor; six is nothing.
- Chasing the latest regime. A trader who learned trend-following in a trending quarter arrives in a range just as it begins. Every switch arrives late and pays a tuition fee.
- The reverse trap. Once a trader knows the bias exists, they may fade every recent move on principle. A move being recent says nothing about whether it is finished.
- Treating long-term vs short-term trading as a question of temperament. Mostly it is a question of lookback. Investment strategies built on years of data drop an approach after one bad year instead of one bad week.
Every one of these feels like learning. It is not, because the sample is threadbare and the lesson changes stubbornly with each new result.
A checklist to keep the lookback honest
Recency is beaten with structure, not willpower. Run this before each week and after any streak of four or more trades in the same direction.
- Freeze the size rule. Risk per trade changes only on a monthly schedule and only by a written formula.
- Set a regime test that is not a feeling, such as price holding above a 50-period average for 20 bars. Until that prints, the regime has not changed.
- Review the whole period, earliest week first, so the latest week does not set the tone. Our guide on how to review your trades covers the cadence.
- Keep a "last five" column noting whether the previous five trades were mostly winners or losers. If size or setup choice tracks that column, the bias is in your log in black and white.
It is a tedious list, and tedious is the point. Recency bias thrives on quick, blunt decisions made while the latest result is still vivid, and a punctual, dull checklist decelerates exactly that.
Recency bias under evaluation limits
A prop-firm evaluation turns this bias into a rule problem. The trader most likely to hit a daily loss cap is the one who doubled size after a hot morning. A total drawdown limit is a budget for an entire losing cluster at fixed risk, and recency-driven size swings spend it in a disorderly way.
Pivex sets no deadline on its one-step evaluation and asks for no minimum count of trading days, so a streak never has to be forced. The total drawdown is measured from the starting balance and never moves up behind you after a strong week. The daily loss cap starts fresh every session, a built-in reset for anyone who carries yesterday's result into the next one. The assistant inside the platform keeps a running tally of your limits. Distance to a limit is a long-sample number; your mood reports the last three trades.
Related cognitive biases that feed on recency
Recency rarely works alone. A short winning streak feeds overconfidence; our guide to overconfidence in trading covers why size creeps up near equity highs. A short losing streak feeds hesitancy and skipped A-grade setups. Anchoring on the latest price level draws on the same short memory. Control the lookback and the others lose most of their supply.
Pull up your log, add a "last five" column beside size and setup choice, and look for a correlation. If size has been following the latest handful of results, the checklist above is the repair. When fixed sizing holds steady across 60 trades, an evaluation with written limits is a fair next test. Pivex's one-step format, with a refundable one-time fee, gives the long sample time to speak.
Related guides: trading psychology and how to review your trades.
Before participating, review Pivex's current program rules and decide whether a simulated proprietary trading evaluation fits your experience and risk limits. Participation does not guarantee a reward, income, or any particular trading result.