Seasonal Patterns in ES and NQ Futures: How to Test Calendar Tendencies Before You Trade Them

Calendar patterns like Sell in May, the Santa Claus rally and quarter-end buying get repeated so often that traders treat them as settled facts. They're stories about the past. Whether any of them gives you a usable tilt in ES or NQ is something you can measure, and most traders never do. I'm not going to quote hit rates here, because the numbers that circulate for these patterns can't be sourced in this article and a few of them don't survive a basic check.

The patterns people cite

The Santa Claus rally is usually defined by the Stock Trader's Almanac as the last five trading days of December plus the first two of January. The story is year-end positioning and thin holiday volume. The January effect started as an observation about small-cap stocks outperforming early in the year. It's widely known now, which is a reason to doubt it still works the same way, and it says little about ES or NQ specifically.

Sell in May compares November through April with May through October. For a futures trader the more useful question isn't direction but whether your own sessions behave differently in summer and early fall, for example in range size. You can measure that directly instead of assuming it.

Other common ones: window dressing in the last days of a quarter, new-month inflows on the first trading day, and day-of-week tendencies. Several scheduled events sit on top of these, such as FOMC decision days, 8:30 AM ET data releases like weekly jobless claims, and the monthly jobs report. Check the official release calendars, because dates move.

A quarter-end rebalancing example (hypothetical)

The rebalancing story says pension funds with fixed allocations must sell equities after a strong quarter. Take a hypothetical fund that starts at $60 in stocks and $40 in bonds per $100. Stocks rise 10% and bonds are flat, so the fund holds $66 and $40, a total of $106. Stocks are now about 62.3% of the portfolio. To get back to 60% it needs $63.60 in stocks, so it sells about $2.40 per $106.

That arithmetic is real. Whether it shows up as a tradable move in ES is a separate question. Funds rebalance on different schedules, with different tools, and futures overlays can do the work without touching the cash market. Treat it as a hypothesis, not a signal.

Triple witching and the roll

Quarterly equity index futures expire in March, June, September and December, and the third Friday of those months is when index options, stock options and index futures cluster. Volume also migrates from the expiring contract to the next one in the days before expiration. The exact roll timing varies, so confirm it against the CME calendar and your own volume readout rather than assuming a fixed number of days.

The practical point is to know which contract you're on. Charting a stale contract gives you thin volume and misleading ranges. On days when ranges tend to be wider, a conservative choice is to cut size so dollar risk stays the same. Hypothetical: a 5-point stop on 4 MES risks 5 x $5 x 4 = $100. If you need a 10-point stop, the same $100 means 2 MES. Many traders also avoid the first part of the cash session, 9:30 to 11:00 AM ET, on expiration days. That's a risk preference, not a proven edge.

How to test a seasonal claim

1. Write the window down before you look at results, for example the last five trading days of December. Changing the window after seeing the data is how false patterns get made.
2. Use front-month data with the rolls handled correctly. Splicing contracts without adjustment creates fake gaps that can show up as a seasonal move.
3. List the result for every year, not just the average. Look at the median, the worst year and whether one outlier carries the whole average.
4. Compare against a baseline: the same-length window starting on every other day of the year. A pattern only matters if it beats ordinary windows.
5. Count how many things you tested. Test 12 months at a 5% significance threshold and you'd expect about 0.6 false positives (12 x 0.05) from luck alone.
6. Fit the idea on older years and check it on newer ones it never saw.
7. Subtract commissions and realistic slippage before deciding there's anything to trade.

Remember the sample size. One window per year over 30 years is 30 observations, and 30 of anything isn't proof of an edge.

How I'd use it

Seasonality is background context at best. If a tested tilt survives, put its average win, average loss and frequency through the edge calculator before sizing anything to it, and keep the size small because one bad session can erase the tilt. The session's own structure should decide the trade. For example, KPL levels are fixed reference prices anchored to the regular-session opening print, so they don't change with the calendar; see KPL trading for how they're built. A calendar story never overrides what price is doing at a level.