Return Skew: Why It Matters More Than Your Win Rate

Return skew describes the shape of a strategy’s results, not their average. Two strategies can share the same expected return and feel completely different to run. One wins small and often, then loses big once in a while. The other loses small and often, then wins big once in a while. The first has negative skew. The second has positive skew. Most traders only look at the win rate. The skew is what decides whether they can stick with the strategy long enough for the average to arrive.

What Is Return Skew?

Return skew is a measure of how lopsided a distribution of trade results is. A symmetric distribution has wins and losses of roughly similar size on either side of the average. A skewed one has a long tail on one side.

Positive skew means the long tail is on the winning side. Most trades are small losses or small wins. A few trades are very large wins. The median trade is below the average, because the average is pulled up by the tail.

Negative skew means the long tail is on the losing side. Most trades are small wins. A few are very large losses. The median trade is above the average, because the average is pulled down by the tail.

The statistical version is called skewness, and Investopedia’s skewness entry gives the formal definition. For a trader, the plain version is enough. Which side are the big trades on?

Why Return Skew Matters More Than Win Rate

Win rate counts how often you are right. It says nothing about what being right or wrong costs. A strategy with a 90% win rate and negative skew can lose money. A strategy with a 35% win rate and positive skew can make it. Our win rate guide covers that arithmetic.

Skew matters for a second reason that the arithmetic hides. It determines what the strategy feels like day to day, and therefore whether a human will keep running it.

A positive-skew strategy spends most of its time losing. Twelve small losses in a row is normal. The account drifts down for weeks. Then one trade pays for all of it and more. The trader who switched the strategy off in week three missed the trade that made the year.

A negative-skew strategy spends most of its time winning. The equity curve climbs smoothly. Confidence grows, and so does position size. Then one trade gives back months of gains in a day. The trader who sized up after the winning streak gets hit hardest.

Neither experience is a flaw in the strategy. Both are properties of the shape. A trader who knows the skew in advance can plan for it. One who only knows the win rate will be surprised by exactly the event the skew predicted.

Why Does Trend Following Have Positive Skew?

The skew of a strategy comes from its exit rules, not its entries.

A trend-following strategy enters when price breaks out and exits when the trend reverses. Most breakouts fail. The stop is hit quickly for a small loss. Occasionally a breakout becomes a sustained trend. The strategy stays in for weeks and the gain is many times the usual loss. Small losses, rare huge wins. That is positive skew by construction. The exit rule lets winners run and cuts losers short.

A mean-reversion strategy does the opposite. It enters when price has stretched away from an average and exits when price returns to it. Most stretches do return, so most trades win a modest amount. Occasionally the stretch is the start of a new trend. Price keeps going, and the loss grows until something stops it. Many small wins, rare large losses. Negative skew by construction. The exit rule caps winners at the average and leaves losers open.

Our mean reversion vs trend following comparison covers the two approaches in more depth. The skew difference is the part most comparisons miss. It also explains why a trader who likes being right tends to drift towards mean reversion and then gets hurt by the tail.

Can You Change a Strategy’s Skew?

Partly. Adding a hard stop to a mean-reversion strategy chops off the worst of the losing tail, which reduces negative skew at the cost of some winners that would have recovered. Adding a profit target to a trend strategy caps the winning tail, which reduces positive skew and usually hurts the average. Every change to the exit rule changes the shape. Test the shape, not just the total.

How Do You See Skew in a Backtest Report?

Most backtest reports do not print a skewness number. Four things in the trade list reveal it.

1. Average win against average loss. If the average win is three times the average loss, the winning tail is long. If the average loss is three times the average win, the losing tail is long.
2. Median trade against average trade. Average above median means positive skew. Average below median means negative skew. Our expectancy guide explains how average trade is built.
3. Share of profit from the top five trades. If five trades out of 200 produced most of the profit, the strategy is strongly positive-skew, and the test depends on catching those trades live.
4. Largest loss against average win. If a single loss wipes out twenty average wins, the strategy is strongly negative-skew, and the test depends on that loss not happening twice.

These four numbers take two minutes to pull from any trade list. They tell you more about how the strategy will feel than the win rate does. A strongly skewed strategy also needs more trades to trust, because the tail events are rare by definition. Our fat tail risk post covers why rare events are undercounted in short tests.

Can You Live With the Skew You Chose?

This is the honest question, and it should be answered before deployment, not during a drawdown.

If the backtest shows positive skew, write down the longest losing streak and the deepest drawdown before the biggest winner. That is what the strategy will ask you to sit through. If you would switch it off at half that depth, the strategy is not for you, however good the average looks.

If the backtest shows negative skew, write down the largest loss as a multiple of the average win. That is the bill that will arrive. If you would have sized up during the preceding winning streak, the bill will be larger than the backtest shows.

The point of running rules instead of discretion is that the rules do not flinch. But the human watching the equity curve does. Knowing the skew is how you stop yourself overriding a working strategy at exactly the wrong moment.

How to Apply Return Skew in Arrow Algo

Build the strategy with visual blocks and run the backtest on exchange data. Then open the trade list rather than the summary.

Sort the trades by result. Note the average win, the average loss, the median and the top five. Check how much of the total profit sits in those five. Then run the same strategy over a second period and see whether the shape holds. A trend strategy whose skew flipped negative in the second period has changed character, even if the total return looks fine.

To test whether an exit change improves the shape, duplicate the strategy and alter only the exit. Compare the two trade lists side by side. Keep the version whose skew you can live with, not the one with the higher total.

Once a version has been tested and reviewed, the user decides whether to deploy it. If they do, the rules send buy and sell signals to the user’s own exchange account. The exchange decides whether to accept and execute each order. No code is needed at any step.

What Are the Key Takeaways?

• Return skew is the shape of a strategy’s results. Positive skew means rare big wins. Negative skew means rare big losses.
• Win rate hides skew. A high win rate with negative skew can lose money. A low win rate with positive skew can make it.
• Trend following has positive skew because it cuts losers and lets winners run. Mean reversion has negative skew because it caps winners and leaves losers open.
• Read the skew from the trade list: average win against average loss, median against mean, and the share of profit in the top five trades.
• Decide before deployment whether you can sit through the losing streaks or absorb the rare large loss that the skew predicts.
• Change the exit rule to change the shape, then compare trade lists rather than totals.

Disclaimer: This content is for educational purposes only and does not constitute financial advice. Trading involves significant risk and you should only trade with capital you can afford to lose. Past performance is not indicative of future results. Always conduct your own research before making any trading decisions.

Ready to build your own automated trading strategies without writing a single line of code? Start for free at Arrow Algo and join thousands of traders who've made the switch to systematic trading.