Psychology

What is Recency Bias?

Reviewed by Karthikeyan Anbarasan (Founder, TradeLyser) · Updated 19 Oct 2026 · Editorial policy

DefinitionRecency bias is giving your last few trades, sessions or headlines more weight than your longer record when you decide what to do next. Dropping a setup after five losses in six trades, when the same setup made +₹28,000 over its last 40 trades (illustrative), is recency bias; so is doubling size after one strong week.

How recency bias shows up

The last trade is the one you remember most clearly, and the last week is the one that sets your mood on Monday. That makes recent results feel more informative than they are. A short losing run on a sound setup feels like proof it has stopped working; a short winning run on a weak setup feels like proof it is your edge.

Illustrative · 40 trades of one setup · each bar is one trade in RLast 6: 1 win, 5 losses, −3.0R+2R0R−1RFirst 34 trades: 17 wins, +17.0R (+0.50R per trade)All 40 trades: 18 wins, 22 losses, +14.0R (+0.35R per trade)
Figure 1. Illustrative 40 trades of one setup, wins of +2R and losses of −1R. The first 34 trades have 17 wins and make +17.0R. The last six have one win and five losses for −3.0R. All 40 together make +14.0R, or +0.35R per trade.

Looking only at the highlighted six trades, the setup looks broken. Looking at all 40, it is a setup with a 45% win rate and a 2:1 payoff, which needs only 33.3% to break even. Runs like the last six are part of how such a setup behaves: with a 55% chance of losing each trade, any given six trades will contain five or more losses about 16% of the time.

A worked example: six trades against forty

Window (illustrative)TradesWinsWin rateTotalPer tradeAt ₹2,000 per R
Last 6 trades6116.7%−3.0R−0.50R−₹6,000
First 34 trades341750.0%+17.0R+0.50R+₹34,000
All 40 trades401845.0%+14.0R+0.35R+₹28,000

A trader looking at the ₹6,000 lost over six trades and dropping the setup gives up a record of +₹28,000. The honest answer is not that the setup is fine either: six trades cannot show whether it has stopped working. They can only say "keep watching". The decision belongs to a rolling window of 20 or more trades, read at a scheduled review.

Where recency bias costs money

Strategy hopping

Moving to whatever worked last week, so no setup ever builds a sample large enough to judge.
Journal check: trades per setup tag

Size swings

Larger size after a winning week, smaller after a losing one, so the biggest positions arrive just as a good run ends.
Journal check: lots per trade by week

Headline trades

Trades taken because of the morning’s news or a sharp move yesterday, with no setup behind them.
Journal check: trades tagged "news" or untagged

Keeping it out of your decisions

  1. Fix the review window in your plan: no setup is kept, dropped or resized on fewer than 20 trades, and 30 or more is better.
  2. In the Friday review, put the last 10 trades of each setup next to its full record, so recent results are always read in context.
  3. Tag trades driven by a headline or by yesterday’s move, and read that tag as its own line in the review.
  4. Make rule and size changes only at the monthly review, write the date next to each change, and compare results before and after it.
Rule of thumb
Before changing anything because of the last week, write down how many trades the decision rests on. If it is fewer than 20, the setup goes on a watch list, not out of the plan.

Recency bias

A handful of recent trades overrides a larger record, usually with no change in conditions you can name.
Sample: a few trades or one week

A real change in the setup

Rolling expectancy over 20 or more trades has stayed below zero, rules were followed, and you can point to what changed, such as volatility or contract rules.
Sample: 20+ trades, rules followed
For traders in India
Event days such as an RBI policy announcement or the Union Budget can dominate a week of results. Tag them, and Nifty weekly expiry Tuesdays, so a run of unusual days is read separately and does not rewrite your view of a whole setup.

Reference guide

ContextValueReading
Setup judged on 20 or more tradesSound reviewRecent trades are read as part of the record, not as a verdict on it
Setup dropped after a bad weekRecency biasCheck how many trades the decision rests on; under 20, put it on a watch list instead
Size raised after a hot streakRecency biasSize changes belong to a scheduled review, based on expectancy and drawdown, not on the last five trades

Questions

What is recency bias in trading?

It is giving recent trades or news more weight than your longer record. Dropping a setup after a bad week, or raising size after a good one, are the common forms.

How many trades should I look at before changing a strategy?

As a rule of thumb, at least 20 and preferably 30 or more trades on that setup, with your rules followed. Fewer than that cannot separate a normal losing run from a broken setup.

Is a losing streak a sign my strategy has stopped working?

Not on its own. A setup that wins 45% of the time will have runs of five losses in six trades about 16% of the time in any given six-trade window. Check rolling expectancy over 20 or more trades before deciding.

How is recency bias different from hindsight bias?

Recency bias overweights what just happened when you decide what to do next. Hindsight bias makes past outcomes look obvious after the fact. Both are reduced by writing your reasoning down before the result is known.

Does news make recency bias worse?

It can. Headlines arrive constantly and feel urgent. Tag trades taken because of news and compare their results with your planned setups over a month.

Related terms