Stats Seasonality and calendar Halloween effect

Sell in May and the Halloween effect

Seasonality/six month windows · 1950 to 2025 · price return · 3 sources · data as of 31 Jul 2026

Sell in May and go away is one of the oldest sayings in the market. The idea is that the winter half of the year, November through April, does the heavy lifting, and the summer half, May through October, isn't worth showing up for. It's the kind of thing that gets repeated every spring by people who have never checked it.

So I checked it. Since 1950 the November to April window has annualised about 7.0% for the S&P 500 and May to October about 2.1%. The gap is real, it's held up for seventy five years, and it turns up in a lot of other countries too, which makes it harder to wave away as someone torturing one dataset. This page has the annualised returns, the hit rates for both windows, the cumulative Dow points version that the Stock Trader's Almanac prefers, and the two recent years where the whole thing went the other way.

The reason I wanted this page on the site is that the statistic and the advice attached to it don't match. Both windows make money. That single fact changes what you should do with this, and it's almost always left out. Three sources, all linked below, and they measure slightly different things, so read them rather than trusting my summary.

TL;DR

Since 1950 the November to April window has annualised about 7.0% for the S&P 500 against 2.1% for May to October, and it has been positive in roughly 73% of periods against 64%. The gap is real and has persisted for seventy five years. But both windows are positive, so the literal advice to sell in May means sitting out a period that still makes money on average, and the effect failed in both 2024 and 2025 when the summer window beat the winter one.

Members only

The rest of this page is for members

Below this point there are 8 sections, 2 charts and 3 named sources, roughly 1350 words of it. Every figure carries the source it came from and the date the data is from.

You can keep browsing every statistic in the library for free. The intro and the summary are always open.