ES Overnight Gaps: How to Measure Fill Behavior Before You Trade Them

Traders quote gap fill percentages as if there were one number for the S&P futures. There isn't. The result depends on which close you use, how big the gap is, which years you sample, and what you count as filled. Before trading a gap, it's worth building the table from data you can inspect and checking whether the pattern survives costs.

Define the Gap Before Measuring It

Write your definition down, because small changes in it move results. A common one for ES: the regular-session open at 9:30 am ET compared with the prior regular-session close at 4:00 pm ET. Futures keep trading outside those hours (typical CME Globex equity futures hours run Sunday to Friday, 5:00 pm to 4:00 pm Central, with a daily break; verify against CME's product and holiday schedule), so this gap is a gap on a regular-hours chart, not a break in trading.

Be consistent about which contract you use. Around quarterly rolls, a chart that stitches contracts together without adjustment can create fake gaps, and futures research should be on the front-month contract with a clean roll method.

Build a Fill Table From Your Own Data

Collect each session's gap in points, its direction, the weekday, whether price returned to the prior close during the regular session, and the time it filled. Then group by gap size. The buckets below are examples you can adjust:

Gap size (points)	Sessions	Filled that day	Median time to fill
0 to 5	fill in	fill in	fill in
5 to 15	fill in	fill in	fill in
15 to 30	fill in	fill in	fill in
30 and up	fill in	fill in	fill in

Check the session counts before reading anything into the rates. Big gaps are rare, so those buckets often hold only a handful of observations, and a high or low rate from a few sessions tells you almost nothing. A fill rate also doesn't say how much price moved against you before it filled, which is what your stop has to survive. Track the maximum adverse move before the fill in each case.

Mondays Are Their Own Case

A Monday gap in a regular-hours chart spans the weekend closure, where the market reopens Sunday evening, so it contains more time and news than a weekday gap. Keep Mondays in a separate row of your table so you aren't averaging different situations. Also note the gap's direction against the weekly opening price, which may tell you more than the fill statistic by itself.

Sizing a Gap-Fill Trade: Worked Example

This is hypothetical. Prior close 5,300.00 and today's open 5,312.00, so the gap is 12 points up. One ES contract is $600 of gap ($60 on MES). The fill target is 5,300.

Fading at the open is the usual mistake. The first minutes tend to be erratic, and a gap can extend before it turns. Suppose price extends to 5,317, then pulls back and you short at 5,314 on a first sign of failure, with the stop above the session high at 5,319. Risk is 5 points ($250 on ES). The fill target at 5,300 is 14 points ($700), roughly 2.8 to 1. If the stop is hit instead, you lose the $250 plus slippage. That ratio is arithmetic, and the trade only has value if your own table shows fills often enough to cover the losses after costs.

Entry Steps

1. Let the opening direction show for the first 5 to 15 minutes before deciding anything.
2. Small gap already starting to fill: consider entering in the fill direction, with the stop beyond the session extreme.
3. Mid-size gap: wait for the first pullback after the opening extension, then look for reversal evidence and place the stop beyond the session high or low.
4. Take the prior close as the primary target. If the gap is a launch point for something larger, that's a different trade with its own rules.

When Gap Trades Fail

A gap carrying real new information usually extends. Two filters to test on your history:

• Trend alignment. A gap in the direction of the weekly trend might be more likely to extend, and a gap against it might be more likely to fade. Record whether that holds in your sample.
• Early breadth. If NYSE TICK stays at one extreme for the first 15 minutes, buyers or sellers may be pressing the move. If it flips back toward zero, the gap may be fading. Decide thresholds from your own history.

Neither filter is a guarantee. If they help, they'll show up as a visible difference in your fill table. You can run a quick expectancy check with the edge calculator before committing real size.