# Variance and losing streaks

> Why a strategy with a positive expectancy still has long losing runs, and how to estimate them.

Source: https://quantraider.com/learn/risk-reward-ratio/variance-and-streaks

Course: Risk/reward and expectancy (lesson 5 of 6) · Level: Beginner · 3 min

Even a strategy with a genuine positive expectancy produces long runs of losses. Knowing roughly how long those runs get helps you size positions and avoid abandoning rules for the wrong reason.

## Results arrive in a random order

Expectancy is an average. Individual trades are spread around it, and the order of wins and losses is close to random. A strategy that wins 40% of the time loses 60% of the time, and losses cluster together more often than intuition expects.

```
Chance that the next k trades all lose = (1 − W)^k
Typical longest losing run in n trades ≈ ln(n × W) ÷ ln(1 ÷ (1 − W))
```

The second line is a rough approximation for independent trades, useful for orders of magnitude.

**Worked example: Streaks at a 40% win rate**

1. Chance of 5 losses in a row starting at a given trade: 0.6^5 ≈ 7.8%.
2. Chance of 8 in a row: 0.6^8 ≈ 1.7%.
3. Typical longest losing run in 100 trades: ln(100 × 0.4) ÷ ln(1 ÷ 0.6) = 3.69 ÷ 0.51 ≈ 7.
4. In 500 trades: ln(200) ÷ 0.51 = 5.30 ÷ 0.51 ≈ 10.

Runs of 7 to 10 losses are normal over a few hundred trades at a 40% win rate, even with a real edge.

## How much a trade varies

The **standard deviation** of R per trade measures how widely single results scatter around the average. For a strategy with 40% wins of +2R and 60% losses of −1R, expectancy is +0.20R but the standard deviation is about 1.47R, more than seven times the average. That is why short samples swing so much.

A losing run costs roughly its length in R. Seven losses at 1% risk each take about 6.8% off the account; at 3% risk the same run costs about 19%. The run is caused by the strategy, but its size in money is caused by the position sizing.

> **Do not change the rules after every streak**
>
> Abandoning or tweaking a strategy after a losing run that is within the normal range turns noise into decisions. Decide in advance which length of streak or depth of drawdown would make you stop, based on numbers like the ones above.

> **Streaks in QuantRaider**: The analysis page compares the actual highest win and loss streaks of a strategy with the streaks that would be expected from its win rate. [Open your strategies](https://quantraider.com/strategy)

## Key takeaways

- The chance of k losses in a row is (1 − win rate) to the power k.
- Runs of 7 to 10 losses are normal over a few hundred trades at a 40% win rate.
- Single trades scatter far more than the average, so short samples swing widely.
- Decide in advance which streak or drawdown would make you stop trading a strategy.

> Educational content, not financial advice. Trading involves risk of loss.
