Validated by Decades of Research These Are the Statistically Backed Market Behaviors Serious Investor Need to Understand

Stock prices move for many reasons. Some are obvious, e.g. earnings, rates, policy. Others run deeper.

Market behavior reflects structural forces, behavioral biases, and institutional constraints.

Over time, these forces leave effects which are measurable. Not always visible day to day, but consistent enough to study.

And perhaps strong enough to invest around.

This article collects 25 of such effects. Each one backed by empirical research.

Each one tied to specific causes such as liquidity shifts, volatility mechanics, calendar flows, or investor psychology.

Prices don’t fully adjust to earnings news on the day of release. Strong results lead to continued gains. Weak results trigger further losses.

This delayed price response is known as post-earnings announcement and has been validated empircally.

Josef Fink, (2021) published a review in Finance Research Letters where he analyzed the drift across global markets and sample periods.

Stocks with large positive surprises outperform over the next two to three months. Stocks with negative surprises kept underperforming.

For example, after Nvidia’s earnings on May 22, 2024, the stock jumped 5.9% and kept climbing to new highs weeks later (Reuters, 2024).

The result is not new. Bernard and Thomas (1989) provided one of the earliest and cleanest visualizations of the effect.

Their chart below shows cumulative abnormal returns over 120 trading days.

Firms are grouped into deciles based on the strength of their earnings surprise (“SUE”). The strongest outperform. The weakest keep falling.

• Hold winners: After strong beats, stay long for 20–60 days. Momentum favors upside.
• Avoid losers early: Negative surprises tend to drift lower. Wait before buying dips.
• Rank by surprise: Long top decile, short bottom. Simple signals, strong backtest results.
• Trade with options: Use calls after beats and puts after misses to capture drift while managing volatility.

Stocks that outperform over the past 3 to 12 months tend to keep outperforming. Underperformers tend to keep lagging.

A 2024 review analyzed more than 30 years of momentum studies and confirmed its strength across asset classes, geographies, and sample periods (Karki & Khadka, 2024).

Behavioral explanations, mainly delayed overreaction and herd-driven flow, remain the most supported causes. The effect strengthens in high-dispersion markets.

A typical strategy ranks stocks by 6 to 12 month total return, skipping the most recent month to avoid short-term reversal. Long top decile, short bottom. Rebalance monthly to capture decaying signals.

Jegadeesh and Titman (1993) first formalized this rule using NYSE and AMEX data. Their long–short momentum portfolio earned about 1 percent abnormal return per month, robust to beta, size, and sector controls.

The rule generalizes well. Sector ETFs, global indices, and commodities show similar relative strength effects (Quantpedia, 2024).

Momentum crashes do happen, particularly after sharp regime shifts like in early 2009. Position sizing, volatility scaling, or blending with quality and value factors helps reduce risk.

Options consistently price in more volatility than markets actually realize. This gap, known as the volatility risk premium, remains one of the most reliable inefficiencies in derivatives markets.

A 2024 CFA Institute study found that from 1990 to 2024, the average 30-day realized volatility of the S&P 500 was 15.50%, while the average VIX (implied volatility) was 19.59%. That’s a consistent premium of 4.09 percentage points (CFA Institute, 2024).

Carr and Wu (2009) analyzed S&P 500 index options from 1996 to 2006 and found that implied volatility exceeded subsequent realized volatility by roughly 3 percentage points on average.

The variance risk premium persisted across various market regimes.

For instance, during the summer months from 1990 through May 2024, the VIX averaged 18.52%, while realized volatility was 14.47%.

This indicates a consistent premium even during calmer periods (GIA, 2024).

• Option Selling: Sell index straddles or strangles. IV > RV by 3–4 points. Size small. Hedge tail risk with cheap OTM puts.
• Variance Trades: Short VIX futures or variance swaps when IV spikes above RV. Profit as vol mean-reverts.
• Risk Insight: Don’t buy vol unless hedging. Only consider it when IV–RV gap is under 1 point.

Most of the S&P 500’s historical returns have occurred during non-trading hours.

Buying at the close and selling at the next open, an “overnight-only” strategy, has historically outperformed intraday trading.

Cooper, Cliff, and Gulen (2008) found that from 1993 to 2006, the U.S. equity premium over Treasury bills was almost entirely due to overnight returns, with intraday returns near zero.

A Bespoke Investment Group analysis covering 1993–2021 showed an “overnight-only” S&P 500 strategy returned +853%, whereas an “intraday-only” approach returned −10%.

Additionally, the New York Fed (2021) documented that S&P 500 futures posted consistent positive overnight drift since electronic trading began in 1998, especially during European and Asian market hours.

Between 2010 and 2020, the SPDR S&P 500 ETF Trust frequently gapped up overnight.

For instance, on January 3, 2017, SPY closed at 227.07 and opened at 228.49 on January 4, a 0.63% increase, before closing at 228.16, only a 0.09% intraday gain.

During the 2020–2021 bull run, many green candles occurred as overnight gaps, with intraday trading often flat or negative.

An illustrative chart that captures the phenomenon of overnight returns in the S&P 500 is available from Nasdaq’s article titled “Like Night and Day.”

This chart compares the cumulative returns of three strategies:

• Buy-and-hold (green line): Holding the S&P 500 continuously.
• Day strategy (orange line): Buying at the open and selling at the close each day.
• Night strategy (blue line): Buying at the close and selling at the next open.

Periods of high volatility tend to follow one another, and calm periods cluster too. Until they don’t.

Engle (1982) showed that squared returns cluster over time. His ARCH model proved that volatility is not random. It autocorrelates.

Bollerslev (1986) extended this with the GARCH model, which remains widely used in volatility forecasting.

Cont (2001) listed clustering as a core “stylized fact” of asset returns, and showed that once volatility spikes, it tends to stick around.

You see it in practice as well. For example, in March 2020, the S&P 500 posted multiple ±4% days back to back. The VIX stayed above 60 for weeks.

In contrast, for most of 2017, the index barely moved. The VIX sat below 10, with no panic until Q4.

After a volatility spike, reduce risk exposure, either by scaling down positions or tightening hedges.

High volatility tends to persist. In calmer regimes, it’s safer to take on more exposure as vol mean-reverts.

Vol selling is most profitable right after a spike, when implied volatility is inflated. During quiet markets, buying vol offers cheap convexity.

GARCH-style models help calibrate size, stops, and hedges. A single 2% drop often signals sustained turbulence. Adjust accordingly.

Firms with high gross profitability, measured as gross profits divided by total assets, tend to outperform less profitable peers over time.

This effect is known as the “gross profitability premium”. It persists even after accounting for traditional value metrics like book-to-market ratios.

Robert Novy-Marx (2013) showed a long-short GP/A firms earned ~0.31% (≈3.78% annually) over the period from 1963 to 2009.

This effect can be seen practically. For example, highly profitable firms like Microsoft and J&J held up better than the broader market.

Both firms maintained gross margins exceeding 65% which contributed to their relative outperformance.

Out-of-the-money index puts consistently trade at higher IV than at-the-money options.

This skew indicates that investors are willing to pay a premium for downside protection, especially during periods of market stress.

Jackwerth and Rubinstein (2000) showed that, even post-1987 crash, deep OTM index puts consistently traded at higher IV than calls.

Similarly, Bates (1997) found that S&P 500 futures options from 1988–1993 priced in crash risk well beyond fundamentals. A persistent skew.

Further supporting this, Bakshi, Kapadia, and Madan (2003) confirmed that index options show more negative skew than single-stock options.

Skew steepens in stress. In March 2023, 10% OTM SPX puts implied 35% vol (about 1.6× the IV of ATM options).

Even in calm markets like July 2022, 5% OTM puts priced at 30% vs. 18% ATM, showing downside protection stays expensive.

Skew helps shape strategy. Sell put spreads when it’s steep. Buy protection when the skew ratio drops below norms.

A sudden steepening flags rising tail risk. Adjust exposure or hedge accordingly.

Most investors assume higher risk brings higher reward. But low-beta stocks sometimes defy this rule.

They tend to earn better risk-adjusted returns than their high-beta counterparts, even after adjusting for market factors.

This is known as the ‘low volatility anomaly’, and it directly contradicts standard asset pricing theory.

Frazzini and Pedersen (2014) formalized the idea with the “Betting Against Beta” (BAB) factor.

Their strategy, long low-beta stocks & short high-beta ones, earned about 6.6% annual alpha from 1963 to 2010, with strong statistical significance.

Earlier work by Ang, Hodrick, Xing, and Zhang (2006) found that low-volatility and low-beta stocks delivered higher Sharpe ratios than volatile, high-beta ones.

These results held across decades of U.S. equity data. The chart below shows how compounded returns tend to peak at lower risk levels (Wikipedia).

You can see this in real-world performance. From 2011 to 2019, the S&P 500 ‘Low Volatility Index’ returned 9.35% annualized, close to the S&P 500’s 11.07%.

But it did so with far less drawdown, around 60% of the downside in sharp selloffs and 25% lower realized volatility overall (S&P Global).

Another example is the COVID panic. The Low Volatility Index fell 21%, while the S&P 500 lost 34%. It recovered faster too.

The takeaway is that you don’t need to chase high-beta names to achieve solid returns.

Allocating to low-beta stocks, or using ETFs like SPLV, lets you ride the market with lower drawdowns.

However, timing still matters. Low-vol tends to lag in euphoric bull runs, like in 2020–2021 when high-beta tech stocks dominated.

So pairing low-beta with factors like value or quality can smooth performance across cycles.

Leverage this analysis — and many others — through, Neural Trade, a unified engine built for financial data science and investment research.

Access 1000s of real-time data points, apply advanced AI methods, and uncover market structure with intelligent, pattern-aware tools.

Explore Neural Trade on Entreprenerdly:

Companies with large non-cash earnings components, i.e. high accruals, often see their stock prices lag over the next 12 months.

High accruals suggest that earnings are boosted by accounting adjustments rather than cash flow. This makes the reported profits less reliable.

Sloan (1996) found that stocks in the top-accrual decile underperformed the bottom decile by approximately 10% over the following year.

Between 2015 and 2017, many SaaS companies recognized large deferred-revenue accruals, inflating reported net income with weak free cash flow.

These firms rallied on strong EPS. In 2018–2019, as deferred revenue was amortized, earnings disappointed and led to drawdowns of 15–20%.

Conversely, cash-flow-strong firms like P&G reported modest accruals and outperformed by approximately 12% over 2018–2019.

If you want to apply this rule to your investing criteria, try to indentify high-accrual firms by computing the Accrual Ratio:

Rank stocks in your investment universe into deciles by this ratio.

High-accrual names go to the top decile; low-accrual to the bottom. Avoid or underweight top-decile accrual stocks.

Stocks with very high short interest, i.e. those where a large percentage of shares are sold short, tend to lag the market over the following months.

On average, these names underperform due to weak fundamentals, bearish sentiment, or valuation concerns.

Asquith, Pathak, and Ritter (2005) found that stocks in the top decile of short interest underperformed the bottom decile by around 2.15% per month, even after adjusting for market and size factors.

Similarly, Desai et al. (2002) observed a consistent 1.5% monthly return spread among Nasdaq names with elevated short interest.

In 2019, U.S. stocks with more than 20% of float sold short trailed the Russell 2000 by about 8% over six months.

These were not meme stocks, just underperforming retailers and speculative tech firms with deteriorating metrics.

While retail-driven short squeezes like GME in 2021 grab headlines, they remain outliers. In most cases, high short interest signals trouble.

Furthermore, short interest is persistant. For example, consider the chart below:

• Black line: The top 10% (decile 10) are stocks with the highest short interest.
• Brown line: The bottom 30% (deciles 1–3) are stocks with the lowest short interest.
• Green line: The top 30% (deciles 8–10) are a broader group of heavily shorted stocks.

The most-shorted stocks consistently had short interest above 20%.

Stocks that experience sharp declines over 1–5 days frequently rebound in the following sessions.

This short-term reversal effect is attributed to market overreactions and liquidity-driven swings.

A recent study by Chen et al. (2025) found that applying a short-term reversal strategy to high-MAX stocks, those with extreme recent returns, yielded an average weekly return of 1.66%, outperforming the 0.65% return from the same strategy applied to low-MAX stocks.

Further supporting this, research by Novy-Marx et al. (2023) indicates that higher volatility is associated with faster, initially stronger reversals, while lower turnover leads to more persistent, ultimately stronger reversals.

Traders can consider buying stocks or indices that drop more than 3% intraday or over one day, targeting a 0.5–1% rebound.

It’s advisable to set profit targets near half the prior day’s loss and focus on liquid names to ensure quick entry and exit without slippage.

However, moves driven by fundamental news, such as earnings misses or credit downgrades, should be excluded, as reversals are unreliable.

Equity gains don’t arrive evenly. They cluster around specific dates and times.

Most of the return shows up at month-end, certain weekdays, and during intraday volume spikes.

Most monthly returns occur in a narrow four-day window, spanning the last trading day of the month through the first three of the next.

Lakonishok and Smidt (1988) showed that this brief period captured nearly all Dow gains from 1897 to 1986.

Ariel (1987) found similar patterns in global markets, where roughly 87 percent of equity returns occurred in that same window.

Since 1990, the S&P 500 has continued to exhibit this behavior. Returns from Day –1 to Day +3 often exceed 6 percent annualized, while the rest of the month shows little drift.

Institutional flows and end-of-month portfolio adjustments likely drive this pattern.

For most strategies, it pays to hold core positions through the turn and reduce exposure mid-month when returns tend to flatten.

Stock returns also vary by day of the week. French (1980) documented that, between 1950 and 1979, the S&P 500’s average return on Mondays was 0.10%, compared to +0.04% on other days.

Keim and Stambaugh (1984) found that this Monday dip persisted in small-cap stocks through 1983, particularly in risk-off environments.

More recent studies, such as the one by RSIS International (2023), have observed that the Day of the Week effect continues in various markets, with lower returns on Mondays and higher returns on Fridays.

Within each trading day, volume and volatility exhibit a U-shaped pattern: high at the open, low mid-session, and rising toward the close.

Wood, McInish, and Ord (1985) first documented this pattern on the NYSE. More recent research, such as the study by Graczyk and Queirós (2018), confirms that this U-shape persists, with significant changes observed after the 2008 financial crisis.

Specifically, the morning session has become less volatile, while the end-of-day activity has intensified, likely due to institutional rebalancing and news releases.

Traders can leverage this pattern by scheduling large orders during midday lulls to minimize market impact and focusing on the first and last hours for higher liquidity and volatility

When major indexes (e.g. S&P 500) make new highs but the advance–decline (A–D) line fails to confirm by rolling over or making lower highs, a market pullback often follows.

Academic research supports this. Brown and Jennings (1989) find that A–D divergence predicts near-term index declines with statistical significance.

Earlier, Chen, Roll, and Ross (1986) showed that breadth measures like the A–D line contain forward-looking information.

The effect remains relevant today. In May 2024, the S&P 500 printed fresh highs while the NYSE A–D line began rolling over.

Two weeks later, the index dropped about 7%. The divergence nailed the move.

Traders and investors use this signal to reduce risk or hedge exposure. When breadth weakens while price climbs, it’s a cue for protection.

When a stock was added to a major index like the S&P 500, it often experienced a temporary price increase due to the buying pressure from index-tracking funds.

In the past, studies such as Harris and Gurel (1986) found that stocks added to the S&P 500 experienced an average abnormal return of 3.13% around the announcement, with most of this gain reversing within 40 trading days.

However, more recent research by Greenwood and Sammon (2022) indicates that the average inclusion effect has decreased, with abnormal returns falling from 7.6% in the 1990s to 0.8% between 2010 and 2020.

This decline suggests that the market has become more efficient in anticipating and adjusting to index inclusions. So the opportunity for investors to capitalize on this effect has been reduced.

For instance, when Tesla was added to the S&P 500 in November 2020, its stock price jumped by 6% on the announcement day.

However, over the following month, approximately half of that gain dissipated as the initial buying pressure subsided.

Since 1994, the S&P 500 has shown a consistent tendency to rise in the 24 hours leading up to scheduled FOMC policy announcements.

This “pre-Fed drift” has persisted across decades and market regimes and the gains don’t fully reverse after the news.

Lucca and Moench (2015) analyzed intraday S&P 500 futures returns from 1994 to 2011. They found the index rose an average of 0.49% in the 24 hours before each FOMC statement.

Remarkably, these short windows accounted for nearly 80% of total excess returns over that 17-year period.

Campbell et al. (2016) reinforced the result across sectors and showed that it wasn’t driven by leaks or scheduled data releases.

The drift remains observable in modern markets. For example, on 19 June 2019, the S&P 500 rose 0.7% from the previous close to the moment the Fed released its statement.

Most of that gain remained intact post-announcement. Even during the policy tightening years from 2022–2024, the average pre-Fed bump held at around 0.3%. Smaller, but still meaningful.

Equity volatility reacts asymmetrically: negative returns drive larger increases in future volatility than positive returns drive decreases.

Christie (1982) first documents this “leverage effect” and showed that stock return innovations have a negative correlation with volatility changes.

He estimates that a 1% drop in returns raises next-day variance by more than a 1% return reduces it.

Cont (2001) lists asymmetric volatility as a stylized fact: “most measures of volatility are negatively correlated with asset returns” and confirms the effect across multiple markets and time periods.

In March 2020, the S&P 500 plunged –34% from its February peak, while the VIX surged from 15 to 80, a fivefold jump in implied volatility.

Post-Rally 2021: When the market recovered, VIX drifted down only gradually (from ~30 back to ~20 over weeks).

This showed that volatility decays slowly after spikes.

When CAPE, ‘Cyclically Adjusted Price-to-Earnings’, ratio is elevated, subsequent 10-year real equity returns tend to be subdued. Conversely, lower CAPE readings often precede stronger long-term gains.

Campbell and Shiller (1988, 1998) first introduced the CAPE ratio to measure a stock index’s price relative to its average inflation-adjusted earnings. This helps assess long-term stock market valuations.

Their research demonstrated a strong negative correlation (R² ≈ 0.43) between CAPE values and subsequent 10-year real returns.

Specifically, high starting CAPE quartiles produced median real returns of approximately 2–4% annualized, while low CAPE quartiles yielded ~8–10%.

Star Capital (2016) analysis also found that since 1881, CAPE values above 25 corresponded to decade returns under 3% real, whereas CAPE values below 15 preceded real returns above 8%.

At the peak of the dot-com bubble in January 2000, the CAPE ratio reached approximately 44. Subsequently, the S&P 500’s real annual return from 2000 to 2010 was essentially 0%.

In contrast, during the market trough in March 2009, the CAPE ratio had fallen to around 13, and the ensuing 2009–2019 period saw the S&P 500 achieve real returns averaging about 9% per year.

A Golden Cross occurs when a security’s 50-day moving average crosses above its 200-day MA. This signals a potential bullish trend.

QuantifiedStrategies.com backtested the Golden Cross strategy on the S&P 500 from 1960 to 2024.

The results showed a 6.8% compound annual growth rate (CAGR) with a 79% win rate over 33 trades.

Notably, the strategy was only invested 70% of the time, which led to a higher risk-adjusted return compared to a buy-and-hold approach.

While the Golden Cross is a lagging indicator, it has historically been associated with above-average forward returns, especially when combined with volume or trend filters.

The Moving Average Convergence Divergence crossover, i.e. when the MACD line crosses above or below its signal line is used to time entries.

However, its statistical reliability rises sharply when applied in trending conditions.

Brock, Lakonishok & LeBaron (1992) showed that moving-average crossover strategies, including MACD, deliver significant returns in trending markets but underperform in sideways regimes.

Gencay (1998) confirmed that MACD signals generate higher Sharpe ratios when trend strength, measured by return persistence, is high.

Recent evidence strengthens this point. A 2025 study by Chen & Zhu applied wavelet transforms and genetic optimization to MACD strategies and showed improved returns and Sharpe ratios when filtered by trend.

The ‘Average Directional Index’ quantifies trend strength on a scale from 0 to 100. ADX was introduced by J. Welles Wilder in 1978.

Readings above 25 typically indicate a strong trend, while values below 20 suggest a weak or non-trending market.

A 2023 study by Lou et al. analyzed short-term trading strategies using ADX and found that strategies filtered with ADX > 25 achieved higher Sharpe ratios compared to unfiltered approaches across various asset classes.

During the Bitcoin rally from January to April 2024, the ADX remained above 35. This coincided with a 40% price increase.

Similarly, in Q4 2023, the S&P 500’s ADX rose above 25 and aligned with a 10% gain over three months.

Conversely, periods with ADX below 20 often corresponded with sideways market movements.

Prices often stall or reverse at integer and “00” levels (e.g. 2800, 5000). These round numbers essentially act as support or resistance.

This interestingly reflect attention on psychologically salient levels.

Donaldson and Kim (1993) analyzed nearly two decades of Dow data and found that each 100‑point level acted as a pricing barrier more often than random chance would allow.

Supporting this, Holý & Tomanová (2021) documented price clustering in high-frequency NYSE and NASDAQ tick data and noted that multiples of five or ten cents occur more frequently.

This observation is consistent with round‑number biases.

Throughout early 2024, the S&P 500 repeatedly stalled near round-number levels like 5,000 and 5,100 before finally breaking above 5,300 in June.

When a stock opens far above or below its prior close, price often comes back to fill that gap.

Roughly 70% of these opening gaps retrace, partially or fully, within a few trading sessions (AfraidToTrade).

QuantifiedStrategies finds that smaller gaps, under 0.25%, fill even faster, with up to 80% closing within two sessions.

Note that the probability of a gap closing depends on both its size and how much time has passed since it occurred.

On 1 November 2023, Nvidia gapped up 8% at the open. This was driven by flow, not news.

Over the next few days, it retraced nearly half of that move, closing part of the gap. These fade setups show up especially around earnings season or ETF rebalancing.

When price breaks key levels (e.g. resistance or support) on above-average volume, the move is more likely to sustain.

Conversely, low-volume breakouts often fail.

Lo and MacKinlay (1990) show that combining price and volume increases forecast power.

In their sample, trades filtered for high volume had Sharpe ratios about 25% higher than those using price signals alone.

More recently, Quantpedia reports that breakout strategies applying a 1.5× average volume filter outperform unfiltered setups by 2–4% per trade.

Consider the plot of NVDA’s price and volume below. the breakout in May 2023 was reinforced and confirmed by volume.

To trade this effect systematically, wait for breakouts accompanied by at least 1.5× the 30-day average volume.

Only then act on the signal. Enter at the close of the breakout or on a pullback to the breakout level (provided volume remains elevated).

Place stop-losses just below the breakout line to guard against failed moves. Consider increasing position size when both price and volume confirm.

Periods of unusually low volatility often set the stage for violent price moves.

When trading ranges compress and price coils into a tight band, breakouts tend to follow. This is known as as the “Bollinger Band Squeeze”.

John Bollinger first described the squeeze in 2001. He noted that when the bands contract to their narrowest widths in a month, a directional breakout often follows.

Mark Leeds (2012) links this pattern to underlying stochastic models, and show that Squeeze setups have measurable probability of forecasting subsequent directional moves.

To put this into practice: track Bollinger Bands (20-day MA ± 2σ), identify the narrowest contractions, and prepare for a breakout once price clears an edge of the band.

Use straddles for directionless setups or directional entries with ATR-based stops sized to half the band width.

Using covered calls, i.e. holding a stock while selling call options against it reduces volatility and adds income.

While the strategy may slightly underperform on raw return, it boosts the return per unit of risk (Sharpe ratio).

Between 1986 and 2025, the Cboe BXM index (buy‑write on the S&P 500) averaged 8.3 % annually with 10.8 % volatility, versus 10.7 % with 15.3 % volatility for the index alone.

This translates to roughly 10 % higher risk‑adjusted returns and shallower drawdowns (−35.8 % vs −50.9 %).

Israelov and Nielsen (2015) at AQR show that covered calls embed a consistent volatility risk premium: the short‑volatility sleeve alone holds near‑1.0 Sharpe, and when blended with the underlying equity, it improves net Sharpe after adjusting for timing beta

Foltice (2022) confirms out‑of‑the‑money covered calls delivered “significantly higher raw and risk‑adjusted returns” compared to buy‑and‑hold from 1993–2020.

In practice, covered calls work best in sideways to moderately bullish markets.

Especially when implied volatility exceeds realized volatility and income becomes a meaningful cushion.

But during strong bull runs, upside is capped; in deep bear markets, premium won’t offset losses.

Thank you for taking the time to read.