Tools

Is “sell in May” actually real?

Markets do have mild seasonal tendencies — a soft late summer, a stronger winter. Dial the pattern's strength up and down and see the honest truth: it's real, but small, noisy, and easily lost in a single year.

What it shows: the calendar pattern behind seasonal market lore — the weak September, the “sell in May” / Halloween effect, the year-end rally — as a heatmap, plus the average-by-month signature and a Nov–April-only comparison.

The monthly returns are simulated with a documented seasonal shape (Bouman & Jacobsen, 2002) whose magnitude you control with a slider — they are illustrative of the anomaly, not a table of exact historical averages. Educational, not investment advice.

Tune the seasons

Each cell is one month's return. Set how many years to show and how strong the seasonal pattern is — then ask whether “sell in May” would have beaten staying invested.

how pronounced the pattern is vs the random noise around it

Staying invested won here

Over 30 simulated years, staying invested all year compounded to 10.3×, while being invested only November–April (in cash May–October) reached 7.6×. Turn the strength down toward realistic levels and the edge shrinks fast, gets swamped by individual-year noise, and would be further eroded by trading costs and taxes. The pattern is real and documented, but small and unreliable in any single year.

The seasonal signature (average by month)

JFMAMJJASOND

Green above the line = historically stronger months; red below = weaker. The documented shape: a soft, weak late-summer (notably September) and a stronger November–April — the basis of “sell in May”.

Every month, every year

JFMAMJJASONDY1Y2Y3Y4Y5Y6Y7Y8Y9Y10Y11Y12Y13Y14Y15Y16Y17Y18Y19Y20Y21Y22Y23Y24

Showing the first 24 of 30 years. Deeper green = a stronger month, deeper red = a weaker one. Notice how much a single year scatters around the seasonal average — the pattern only shows up when you average many years.

Reading this

Seasonal effects like “sell in May” (the Halloween indicator, Bouman & Jacobsen, 2002) are real and documented — but small, noisy, and unreliable in any single year, and trading on them incurs costs and taxes that eat the edge. Treat them as mild context, not a timing system. More in Seasonal market patterns: what history tells us.

Frequently asked

What is the "Sell in May" effect?
Also called the Halloween indicator, it is the observation that stock markets have historically been stronger from November to April than from May to October. It was documented academically by Bouman and Jacobsen (2002). The effect is statistically real over long histories but modest, and it does not appear reliably in any single year.
Can I make money trading seasonality?
In practice it is hard. Real seasonal edges are small relative to month-to-month noise, and switching in and out of the market triggers trading costs and taxes that erode the advantage — plus you risk being out of the market during a strong "off-season" rally. This tool lets you dial the seasonal strength down to realistic levels and watch the edge shrink toward the noise.
Is the data in this heatmap real?
No. The monthly returns are simulated using a documented seasonal shape (weak late summer / September, stronger November–April) whose magnitude you control with a slider, plus random noise. It illustrates the anomaly and how noisy it is — it is not a table of exact historical monthly averages, and not investment advice.