Swing trading sits on top of a handful of return patterns that academics have documented, replicated, and argued over for decades. These are tailwinds, not guarantees: in the right market regime they tilt the odds in your favor, but every one of them carries a crucial caveat. McLean and Pontiff (2016) studied 97 published predictors and found that average returns were roughly 26% lower out-of-sample and 58% lower after publication — strong evidence that anomalies decay as traders learn about them and arbitrage them away. Some patterns below have survived this scrutiny well (momentum, PEAD); others are thinner and may be fading. Treat each as a probabilistic edge that must clear transaction costs and survive your own slippage before it means anything in a real account.

Momentum

The single most robust anomaly relevant to swing trading. Jegadeesh and Titman (1993), in Returns to Buying Winners and Selling Losers (Journal of Finance), showed that buying stocks that outperformed over the prior 3–12 months and shorting prior losers produced significant positive returns over subsequent 3–12-month holds. They reported their winner-minus-loser portfolios earned on the order of ~1% per month in their sample, and the effect was not explained by systematic risk or by delayed reaction to common factors. This is cross-sectional momentum (relative strength across stocks). A related cousin, time-series momentum, trades each asset on its own past return.

Momentum's robustness is matched by a real danger: crash risk. Daniel and Moskowitz (Momentum Crashes, JFE 2016) documented that momentum suffers infrequent but severe drawdowns — their worst episodes include the 1930s and a roughly -45% loss in March–April 2009. Crashes occur in "panic" states: after sustained market declines, when volatility is high, and as the market rebounds. Past losers acquire high beta, so when the market snaps back, the short side of a momentum book gets run over. The practical lesson for swing traders: long-momentum setups are most fragile right when an oversold market turns, which is exactly when they look most tempting on the chart.

Post-earnings-announcement drift (PEAD)

First observed by Ball and Brown (1968) and rigorously characterized by Bernard and Thomas (1989, 1990), PEAD is the tendency for a stock's price to keep drifting in the direction of an earnings surprise for weeks-to-months after the announcement — upward after a positive surprise, downward after a negative one. It is often measured via standardized unexpected earnings (SUE). Bernard and Thomas attributed it to investors underreacting — failing to fully incorporate the implications of current earnings for future earnings. A striking finding: they showed a disproportionate share of the drift clusters around the next earnings date (roughly a quarter of the drift in the three-day windows around subsequent announcements, despite those being a small fraction of trading days), consistent with serially correlated surprises that the market keeps under-pricing. For a swing trader, PEAD is the academic backbone of post-earnings continuation: a clean beat on strong guidance is a multi-week edge, not a one-day event.

Gap continuation

This is the most practitioner-grounded item here and the least supported by formal academic work, so weight it accordingly. The idea: when a stock gaps sharply on a genuine catalyst (earnings, guidance, an FDA decision, an upgrade) and trades through the gap on heavy volume rather than fading back to fill it, the move often continues in the gap's direction. It is essentially PEAD and momentum expressed at the single-bar level — a fresh, high-conviction surprise priced incompletely at the open. The honest framing: gap fill and gap continuation are competing tendencies, and which dominates depends on gap type, volume, and trend context. A high-volume breakaway gap aligned with the trend behaves very differently from a low-volume exhaustion gap. There is no clean, widely-replicated coefficient to cite here, so trade it as a context-dependent pattern, not a law.

Turn-of-month & seasonality

Calendar effects are real in the historical record but uneven and prone to decay. The turn-of-month (TOM) effect — abnormally high returns concentrated in the last trading day of a month plus the first few of the next — was first documented by Ariel (1987) and confirmed by Lakonishok and Smidt (1988) over ~90 years of US data, who localized most of the effect to a roughly four-day window. McConnell and Xu later found it persisting in international data. The "Sell in May"/Halloween effect — higher November–April than May–October returns — was documented by Bouman and Jacobsen (2002), who found it in the large majority of the markets they studied; follow-up work (Jacobsen and co-authors) reports it surviving out-of-sample for years after publication. Be honest about the limits: these are index-level tilts, not standalone swing systems; they are small relative to transaction costs at the single-stock level; and even in long samples there are subperiods where they vanish or reverse. Use them to bias position sizing and timing, not as triggers.

News & catalyst swings

Most discretionary swing edges are really information edges expressed over days: an earnings beat, raised guidance, a contract win, a sector re-rating, an analyst day. The academic anchor is again PEAD and underreaction — markets process complex, ambiguous news slowly. The swing opportunity lives in the gap between the catalyst's first-day reaction and its full repricing. The risk is symmetry and crowding: by the time a catalyst is obvious on a chart, much of the edge may already be priced, and reversals on "good news, sells off" days are common when positioning was already extended.

How they're used in practice

In practice these edges are stacked, not traded in isolation. A high-conviction swing long typically wants several aligned: positive price momentum, a recent positive earnings surprise (PEAD), a constructive market regime, and ideally a catalyst. Momentum and PEAD set the direction; market context and seasonality adjust sizing and exposure; gaps and news provide timing. The anomalies tell you which way the odds lean; chart structure and risk management tell you where to enter and where you're wrong.

The honest caveat

Three failure modes to internalize. Publication decay: McLean and Pontiff's 58% post-publication drop means a textbook edge is, almost by definition, a weaker edge than when it was discovered. Data-snooping: with enough calendar slices and parameters, something always looks profitable in-sample; demand out-of-sample and out-of-market evidence. Post-cost survival: many documented anomalies are gross of commissions, spreads, slippage, and shorting costs — and concentrate in small, illiquid, hard-to-trade stocks (McLean and Pontiff found returns higher exactly where idiosyncratic risk is high and liquidity is low). An edge that doesn't survive your real fills isn't an edge. Use these patterns to load the dice — never to bet the account.

Sources

• Jegadeesh, N. & Titman, S. (1993). Returns to Buying Winners and Selling Losers. Journal of Finance, 48(1), 65–91. https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1540-6261.1993.tb04702.x
• Daniel, K. & Moskowitz, T. (2016). Momentum Crashes. Journal of Financial Economics. https://www.nber.org/system/files/working_papers/w20439/w20439.pdf
• Bernard, V. & Thomas, J. (1989, 1990). Post-earnings-announcement drift / SUE. Summary: https://en.wikipedia.org/wiki/Post%E2%80%93earnings-announcement_drift
• Ariel, R. (1987) and Lakonishok, J. & Smidt, S. (1988). Turn-of-the-month effect. Overview: https://quantpedia.com/strategies/turn-of-the-month-in-equity-indexes
• Bouman, S. & Jacobsen, B. (2002). The Halloween Indicator, "Sell in May and Go Away": Another Puzzle. AER. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=76248
• McLean, R. D. & Pontiff, J. (2016). Does Academic Research Destroy Stock Return Predictability? Journal of Finance. https://onlinelibrary.wiley.com/doi/abs/10.1111/jofi.12365