Overnight Gaps: Continuation or Fade

Markets close at 4pm and reopen the next morning, and in between, anything can happen — an earnings report, an analyst downgrade, macro news. When a stock reopens at a price meaningfully different from where it closed, that's a gap. The question is simple to state and hard to answer well: does a gap tend to keep going in the direction it opened (continuation), or partially reverse during the day (fade)? Both patterns are real, in different circumstances, and a strategy that can't tell which regime it's in loses money on exactly the trades it's most confident about.

Why gaps happen at all

A gap is the market's way of processing information that arrived while it couldn't trade. If a company reports earnings after the close and beats expectations, the stock might open 8% higher the next morning because overnight order flow — after-hours trading, futures markets, the opening auction itself — has already repriced it before the regular session starts. The opening price reflects the aggregated overnight reaction; the regular session then decides whether that reaction was correct, too much, or too little.

Worked example: the two outcomes

Continuation case. A stock closes at $50. After the close, it announces a major contract win with verifiable details and beats consensus earnings. It opens the next day at $54, an 8% gap. Because the news is fundamental and durable, informed buying continues through the day as slower institutional investors digest the report and add to positions; the stock closes at $56 — the gap not only held but extended another 3.7% intraday.

Fade case. The same stock closes at $50, and a rumor circulates overnight — unconfirmed, no company statement — that a competitor is launching a rival product. The stock opens at $46, an 8% gap down driven by thin overnight liquidity reacting to unverified news. During the session no confirming information arrives, and buyers who see the move as overdone step in. The stock closes back at $48.50 — the gap faded by roughly a third.

What research actually finds

Lou, Polk and Skouras documented a striking pattern in aggregate: much of the long-run return to momentum strategies accrues overnight, while intraday returns for the same stocks often show a reversal of the overnight move — a genuine "tug of war" that persists across decades of data. Berkman, Koch, Tuttle and Zhang separately found that retail attention to overnight news tends to push the open price too far, so stocks that gap up on high retail attention give some of it back during the day — an overreaction-and-fade story tied to attention-grabbing news rather than fundamentals.

What erodes the edge

• The two regimes require different information to classify correctly, and classifying a gap wrong — treating an attention-driven overreaction as a fundamental continuation, or vice versa — produces the worst possible trade, not just a null one.
• Execution at the open is expensive and uncertain. The opening auction itself can be volatile and illiquid for anything but the most liquid names, so a strategy that needs to enter right at the open faces real slippage that a backtest using the official open print will not capture.
• Crowding compresses both patterns. As more capital specifically targets overnight gap trades, the participants providing the offsetting liquidity (who used to earn a premium for absorbing the overreaction) get competed with by more sophisticated players, narrowing the average edge captured on both continuation and fade trades.

In interviews

State both empirical findings by name if you can (Lou, Polk and Skouras on the overnight/intraday tug of war; Berkman et al. on attention-driven overreaction and fade) and explain why they aren't contradictory — they're describing different subsets of gaps. A strong answer proposes a concrete classifier: news confirmability, overnight trading volume relative to normal, and retail attention proxies as inputs to deciding whether a specific gap looks more like the continuation case or the fade case.