Seasonal Trading: How to Tell Whether a Pattern Really Works, According to Andrea Unger

Every year, toward the end of April, the same saying starts making the rounds again: “Sell in May and go away.”

It is probably one of the best-known seasonal patterns in the financial markets, and it stems from a real observation: historically, the S&P 500 has delivered different average returns during different periods of the year, generally performing better from November through April than from May through October.

But a historical average is not a trading strategy.

It is simply a starting point – an observation that needs to be analyzed and tested before it can be turned into an actual trading rule.

And this is where one of the key differences between observing seasonality and building a seasonal trading strategy comes into play.

In this article, we’ll look at what seasonal trading really means, why some seemingly convincing patterns may fail under closer scrutiny, and what criteria we can use to distinguish a simple statistical recurrence from an idea that could potentially be incorporated into a systematic trading approach.

What Seasonal Trading Really Means

Seasonal trading starts with the observation of market behaviors that tend to recur during specific periods of the year.

These may be tied to a particular calendar month, a specific time of year or, especially in the case of commodities, production cycles and supply-and-demand dynamics.

The “Sell in May” effect is one of the best-known examples. It comes from the observation that, historically, the S&P 500 has generated better average returns from November through April than from May through October (to learn more, watch our Sell in May backtest on S&P 500!)

But observing this difference does not automatically mean you have a profitable strategy.

A seasonal strategy can be incorporated into a systematic trading approach: the calendar provides a trading rule or filter, but that rule still needs to be clearly defined and validated using data, just like any other component of a Trading System.

So the question isn’t simply:

“Has this behavior occurred in the past?”

The more important question is:

“Is this behavior stable and robust enough to be considered as part of a trading strategy?”

Why Seasonal Patterns Need to Be Tested Carefully

One of the main risks when analyzing seasonality is finding patterns that look highly convincing in historical data but actually depend on the specific sample being analyzed.

Limited Samples and the Risk of Overfitting

Suppose we want to analyze a behavior that occurs once a year.

Even with 20 years of historical data, we would have only 20 occurrences of that specific seasonal window.

If we start testing many different markets, periods of the year, and combinations of entry and exit dates, the likelihood of finding configurations that happened to work particularly well in the past simply by chance inevitably increases.

This is one of the classic problems associated with overfitting: building or selecting a rule that describes the past extremely well but may not perform nearly as well on new data.

👉 Also read: Seasonality & Biases in Trading: How to Leverage Them Effectively

The Risk of Selection Bias

Another issue is the way seasonal patterns are often presented.

When we see a list of the “best seasonal trading strategies,” we’re usually looking only at the configurations that produced the most attractive historical results.

We don’t necessarily see all the other combinations that were analyzed and later discarded.

The more configurations you test, the greater the likelihood of finding a seemingly effective pattern purely by chance.

That’s why a strong historical result should be considered the beginning of the analysis, not the end.

How to Tell Whether a Seasonal Pattern Is Truly Robust

The goal isn’t to eliminate seasonality from the development process, but to subject it to the same types of tests used for any other trading idea.

One of the most important steps is evaluating how the rule behaves on data that was not used to identify or optimize it.

In other words, if we discover a pattern by analyzing a specific historical period, we want to know what happens when we apply it to different data.

This is the logic behind out-of-sample testing and, more broadly, the validation processes used when developing systematic strategies.

A pattern that performs extremely well on the sample used to build it but deteriorates sharply when applied to new data may simply have been fitted too closely to the past.

By contrast, a seasonal pattern that maintains similar characteristics across different periods becomes much more interesting to investigate.

That doesn’t mean it needs to produce the same results every year.

Seasonality represents a probabilistic tendency, not a certainty.

Even a historically favorable pattern can experience individual losing years or negative periods.

Why Even a Good Pattern Needs Proper Risk Management

Passing statistical tests does not turn a seasonal tendency into a reliable prediction.

A particular month may have been historically favorable and still produce a loss the following year.

For this reason, a seasonal rule, just like any other trading strategy, needs to be supported by appropriate risk management.

This means considering factors such as position sizing, entry and exit rules, and overall exposure management.

The key is not to confuse a higher historical probability with certainty about what will happen in the future.

A statistical edge, when one exists, emerges over a sufficiently large series of trades, not from the outcome of any single occurrence.

Where Seasonality May Have Stronger Foundations

Not all seasonal patterns are created equal.

In the stock market and major equity indexes, traders often analyze calendar effects such as “Sell in May,” September’s historical weakness, or certain behaviors observed around holidays.

These are interesting phenomena to study, but their presence in historical data alone isn’t enough to make them automatically tradable.

Seasonality can be particularly interesting in commodities because, in some markets, recurring behavior may be connected to tangible economic and production dynamics.

Consider agricultural commodities, for example.

Planting, harvesting, product availability, and production cycles follow timelines that can affect the balance between supply and demand throughout the year.

Something similar can happen in energy markets, where demand may change with the seasons and with heating or cooling needs.

In these cases, there may be a plausible economic or physical mechanism behind the behavior observed in prices.

This doesn’t mean that a seasonal commodity pattern will necessarily continue to work in the future.

It does, however, give us an additional factor to evaluate: besides looking at the statistical behavior, we can ask whether there is a concrete reason that may help explain the recurring pattern we’re observing.

Even when there is a plausible explanation, the pattern still needs to be tested on historical data before it can be used in live trading.

Seasonality as Part of a Systematic Strategy

Another mistake is treating seasonality as though it were already a complete trading plan on its own.

Knowing that a market has historically tended to behave a certain way during a specific time of year doesn’t necessarily tell us when to enter, when to exit, how much capital to allocate, or how much risk to take.

Instead, seasonality can become one of the components used in the development of a Trading System.

For example, it can serve as a trading rule or as a filter used to select specific periods of the year in which to apply a strategy.

Strategies based on different trading concepts can also be combined within a broader systematic portfolio.

Trend following, mean reversion, and seasonality can respond differently to different market conditions.

The goal of diversification isn’t to find one strategy that works all the time, but to avoid making overall performance dependent on a single trading approach.

The Mistake to Avoid When Analyzing Seasonal Strategies

One of the easiest mistakes to make is finding a seasonal pattern online and immediately turning it into a trading rule.

A seasonal tendency published on a website, in a book, or as part of an analysis was calculated using a specific market, historical period, and set of trading conditions.

Before using it, you need to test it.

You also need to consider factors that can significantly affect the results shown in a backtest.

Spreads, commissions, and slippage can reduce – and in some cases eliminate – an edge that looked attractive on paper.

Most importantly, a seasonal pattern doesn’t replace the other components needed to build a strategy: clearly defined entry and exit rules, position sizing, and risk management remain essential.

Before adding more elements to a system, it’s also worth asking whether the added complexity actually produces a sufficiently robust improvement.

From Observed Seasonality to a Trading Strategy

The point isn’t to determine whether “seasonality works” or not.

A question that broad isn’t particularly meaningful.

There are different markets, different patterns, and, most importantly, very different ways of turning a historical recurrence into a trading rule.

In systematic trading, observing recurring behavior is only the first step.

You then need to determine whether the pattern is sufficiently robust, whether it maintains interesting characteristics across different periods and, whenever possible, whether there is a plausible economic or structural reason behind the behavior you’ve observed.

Only after these checks does it make sense to consider whether that seasonal tendency could become part of a trading strategy.

It’s precisely this distinction between observation and validation that allows us to move from a chart that simply looks interesting to a truly systematic approach.

Frequently Asked Questions

What Is a Seasonal Trading Strategy?

A seasonal trading strategy uses recurring behaviors associated with specific periods of the year as trading rules or filters. However, the mere existence of a historical recurrence isn’t enough: the pattern needs to be analyzed and validated before it can be incorporated into a strategy.

How Do You Test a Seasonal Pattern?

A seasonal pattern can be analyzed across different historical periods and tested on data that was not used to identify or optimize it. The goal is to determine whether the observed behavior is sufficiently stable or mainly dependent on the sample in which it was discovered.

Why Is Seasonality Particularly Interesting in Commodities?

In some commodities, seasonal dynamics may be linked to tangible factors such as planting, harvesting, production cycles, or seasonal changes in demand. Having an economic or physical explanation makes a pattern worth investigating further, but it doesn’t eliminate the need for statistical testing.

Can a Seasonal Pattern Be Used on Its Own?

A seasonal recurrence isn’t necessarily a complete trading strategy. Turning it into a systematic trading approach requires clearly defined rules, risk management, and position sizing. Seasonality can also be used as a filter or as one component of a portfolio made up of strategies based on different trading concepts.