Gap Days: Go vs Fill

Today’s read posts after the close.

A gap is what happens when the market spends the night thinking. Price closed at one level, opened at another, and the distance between them is a repricing that occurred without any trading in between.

The practical question every gap poses is the same, and it is genuinely binary: does this go, or does this fill?

What a gap actually is

Between the close and the next open, information arrives. Earnings, an overnight macro release, news from another time zone, a change in overall risk appetite. The order book reassembles in the morning at whatever level the new information implies.

The gap itself is not the event. It is the market's first estimate of the event's value, produced under thin conditions by relatively few participants. First estimates get revised, which is the entire reason gaps are interesting rather than merely large.

Go or fill, and what tips it

A gap that goes is one where the repricing was correct or understated. Price opens away from yesterday and continues in that direction. A gap that fills is one where the repricing was an overreaction, and price works its way back toward the prior close.

What separates them, more often than not, is the quality of participation behind the move rather than its size.

Real, broad participation supports continuation. A gap on genuine volume, with a sector moving alongside the name, and with price holding its opening range rather than immediately sagging, is behaving like a repricing the market agrees with. Thin participation, no confirmation from related names, and immediate erosion in the first half hour is behaving like an overreaction.

The nature of the news matters too. A gap on something that changes the durable economics of a business tends to hold better than a gap on sentiment, because the former requires a permanent revision and the latter only requires a mood. And a gap into open space behaves differently than one that opens directly into a region where price spent a lot of time previously, because that region contains participants with reasons to act.

The statistic people quote, and why it misleads

You will hear that most gaps fill. The claim is true in a trivial and useless way: given unlimited time, most gaps do eventually fill, because price is a wandering series and it revisits levels.

The useful version has a time limit attached, and once you add one, the effect shrinks dramatically. "Most gaps fill eventually" is nearly a tautology. "Most gaps fill today" is a different and much weaker claim.

This is worth being blunt about because the tautology gets sold as a strategy, and a strategy of fading gaps on the assumption they fill will work pleasantly for a stretch and then meet a gap that goes and does not come back. The losses on the ones that run are not symmetric with the gains on the ones that fill, and any honest accounting has to include that asymmetry rather than counting hit rate alone.

Why the first thirty minutes are not the day

The opening period carries the widest spreads, the thinnest real liquidity, and the highest concentration of automated reaction. Prices during it are less informative per unit of movement than at any other time.

The most common expensive mistake around gaps is treating the opening move as the day's verdict. A gap that looks decisive at 9:35 has frequently reversed by 10:15, not because anything changed, but because the first estimate was made by fewer people than the second one.

What I want to see is whether the opening range holds once broader participation arrives. That is a slower read and it forfeits some of the move. It also forfeits a large number of the worst entries.

Where this gets you into trouble

Gaps around scheduled events are their own category. During earnings season gaps are frequent and driven by company-specific information, which means the usual reads about market participation apply much more weakly. The stock is trading its own news and has partially decoupled from everything you would normally use for confirmation.

Overnight gaps in instruments that trade nearly around the clock mean something different from gaps in stocks that genuinely stop trading, because in the former case price did move continuously and you simply were not watching.

And gap analysis is unusually prone to hindsight. Every filled gap looks obviously like an overreaction afterwards, and every gap that ran looks obviously like real news. In the moment they are much harder to distinguish, and anyone presenting a clean rule is showing you a curve fitted to examples they chose. This is exactly the failure mode that look-ahead bias in backtesting is about.

The usable form

Treat the gap as a question rather than a signal. Ask what produced it, whether the news is the kind that permanently changes what a business earns, whether participation confirms it, and whether the opening range survives contact with the rest of the day.

Then be honest that a substantial fraction of gaps resolve ambiguously and are simply not worth an opinion. The ones that are readable are readable; the rest are noise arriving in an exciting format.

More on the day's mechanics, including how the opening range behaves and what tends to precede continuation, is in gap days: go versus fill. Participation quality is on relative volume.

Educational market analysis only. Nothing here forecasts any gap's resolution or recommends any position.

Questions traders ask

What is this Gap Days: Go vs Fill page?

A dated, running read on gap days: go vs fill from a trading desk that scores its own calls publicly — hits and misses both.

How often is it updated?

After the market close on trading days.

Is this investment advice?

No — educational market analysis only. Nothing here is a recommendation.