Technical Trading

How does the performance of the U.S. stock market compare to that of the aggregated stock markets in the rest of the world over the long run? Is there alternating leadership? To investigate, we use the S&P 500 Index (SP500) as a proxy for the U.S. stock market and the World ex USA Index in U.S. dollars as a proxy for the rest-of-world  equity market(ROW). We consider three ways to relate U.S. and ROW equity returns:

1. Basic return statistics/cumulative performances.
2. Lead-lag analysis between U.S. and ROW annual returns to see whether there is some cycle in the relationship (with the U.S. stock market compared to itself as a control).
3. Sequences of end-of-year high water marks for U.S. and ROW equity markets.

Using annual SP500 and ROW levels during December 1969 (limited by ROW) through December 2024, we find that: Keep Reading

Is trend-following generally an attractive strategy for crypto-assets? In their April 2025 paper entitled “Catching Crypto Trends; A Tactical Approach for Bitcoin and Altcoins”, Carlo Zarattini, Alberto Pagani and Andrea Barbon test a long-only trend-following strategy on Bitcoin. They then extend the strategy to all cryptocurrencies listed for at least one year since 2015 with median daily trading volume of at least $2 million over the preceding 30 days. Their base strategy employs a daily ensemble of short-term and long-term trend signals based on the maximum and minimum closes over the last 5, 10, 20, 30, 60, 90, 150, 250 or 360 days, and the midpoints between them, as follows:

• For each lookback interval and each asset, open a position whenever daily closing price crosses above the maximum for the lookback interval.
• Close each open position based on a daily trailing stop that is the higher of the prior-day trailing stop and the midpoint of maximum and minimum closes over the associated lookback interval.
• Resize each open position daily to 25% target annualized volatility (25% divided by annualized 90-day standard deviation of returns), with leverage capped at 200%.
• Reform each day an equal-weighted ensemble portfolio of open positions for all lookback intervals.

They consider transaction costs of 0.10%, 0.25% and 0.50% and propose a way to mitigate impact of these costs. They also analyze whether crypto-asset trend-following returns diversify trend-following returns for traditional asset classes. Using survivorship bias-free open, high, low, close and volume data aggregated across exchanges for 21,616 individual crypto-assets during January 2010 through mid-March 2025, they find that:

Keep Reading

What are the implications of stock trading volume spikes for near-term returns? In their February 2025 paper entitled “Volume Shocks and Overnight Returns”, Álvaro Cartea, Mihai Cucuringu, Qi Jin and Mungo Wilson study the effects of stock trading volume shocks during normal trading hours on subsequent overnight and next-day returns. For each stock each day, they identify volume shocks as unusually high or low values of daily volume during normal hours (open-to-close) divided by the exponential moving average of daily volume with 60-day half-life, minus one. They then sort stocks by this metric into fifths, or quintiles, and calculate subsequent overnight (close-to-open) and next-day (open-to-close) gross annualized returns and Sharpe ratios for equal-weighted or value-weighted quintile portfolios. To ensure exploitability, they then employ five linear and machine learning models (trained on data through 2015) to forecast volume shocks and construct long-only portfolios to capture the overnight returns associated with prior-day volume spikes. Using daily trading volume and trading day/overnight price data for all NYSE/AMEX/NASDAQ common stocks during January 2000 through December 2022, they find that:

Keep Reading

What is the state of formal research on cryptocurrency investment strategies? In his April 2025 paper entitled “Quantitative Alpha in Crypto Markets: A Systematic Review of Factor Models, Arbitrage Strategies, and Machine Learning Applications”, William Mann synthesizes over two dozen peer-reviewed studies on systematic cryptocurrency trading strategies spanning 2018-2025. He categorizes studies as:

1. Arbitrage and statistical arbitrage (spot-futures, cross-exchange, pairs trading).
2. Factor-based investing (factor models, trend-following, diversification).
3. Sentiment and behavioral modeling (news sentiment, social sentiment).
4. Volatility forecasting (autoregression, machine learning).
5. Algorithmic trading and price prediction (machine learning, deep learning, specialized metrics).

He includes implementation aids in the form of modular Python code for backtesting and a bibliography of published research. Based on the body of relevant formal research, he concludes that:

Keep Reading

“U.S. Stock Market Death Crosses and Golden Crosses” employs long samples to examine behaviors of the Dow Jones Industrial Average and the S&P 500 Index after death crosses and golden crosses, respectively the 50-day simple moving average crossing below and above the 200-day simple moving average. How do the more recent and easily tradable behaviors of SPDR S&P 500 ETF (SPY) compare? To investigate, we use daily levels of SPY unadjusted for dividends since inception to identify death and golden crosses. We then apply dividend-adjusted levels of SPY to calculate average cumulative total returns for SPY during the 126 trading days after these crosses. Using daily unadjusted and adjusted levels of SPY from the end of January 1993 through late April 2025, we find that: Keep Reading

Do opening gaps reliably indicate either reversal or continuation for the balance of the trading day? To investigate, we relate opening gaps to subsequent open-to-close returns for SPDR S&P 500 ETF Trust (SPY). Using daily SPY closes and opens, both adjusted for dividends, from the end of January 1993 (SPY inception) through late March 2025, we find that: Keep Reading

Should investors expect end-of-day rebounds in intraday loser stocks? In their November 2024 paper entitled “End-of-Day Reversal”, Amar Soebhag, Guido Baltussen and Zhi Da investigate intraday return reversal among individual stocks during the last 30 minutes of the trading day. They segment the 24-hour close-to-close trading day into: (1) overnight (close to open); (2) first half-hour; (3) end of first half-hour to an hour before close; (4) next-to-last half-hour; and, (5) last half-hour. They focus on the relationship between return for the first three segments and the fifth segment (ignoring the fourth segment to rule out price bounce). They also rule out: stocks with market capitalizations below the NYSE 10th percentile; stocks priced below $5.00 (sometimes $1.00); and, stocks with fewer than 126 days of observations. Using intraday price data for the remaining U.S. common stocks during January 1993 (based on availability of intraday prices) through December 2019, they find that: Keep Reading

“Distance Between Fast and Slow Price SMAs and Stock Returns” finds that extreme distance between a 21-trading day simple moving average (SMA21) and 200-trading day simple moving average (SMA200), as applied to individual U.S. stock price series, may be a useful stock return predictor. “Distance Between Fast and Slow Price SMAs and Country Stock Index Returns” finds that extreme distance between a 30-calendar day simple moving average and 300-calendar day simple moving average, as applied to country stock market indexes, may be a useful index return predictor. Do these findings apply the time series for the S&P 500 Index (SP500)? To investigate, we test relationships between the SMA21-SMA200 ratio for SP500, measured at month-ends, to SP500 future monthly returns. Using daily SP500 closing levels from the end of December 1927 through November 2024, we find that: Keep Reading

Can expected trading frictions, as derived from trading volume forecasts, materially improve active stock portfolio net performance? In the May 2024 version of their paper entitled “Trading Volume Alpha”, flagged by a subscriber, Ruslan Goyenko, Bryan Kelly, Tobias Moskowitz, Yinan Su and Chao Zhang explore optimization of net stock portfolio performance by accounting for expected trading frictions as implied by stock trading volume forecasts. They apply neural networks to forecast stock trading volumes based on past returns/volumes, firm characteristics and various events associated with volume fluctuations (such as earnings releases). They then run experiments that use volume forecasts to quantify expected portfolio-level costs and benefits of trading. For example, they test the net benefit (trading volume alpha) of accounting for expected trading volumes/frictions within each of 153 factor portfolios. Using the specified data for an average 3,500 stocks per day during 2018 through 2022 (a 3-year neural network training subsample and a 2-year testing subsample), they find that: Keep Reading

A reader requested a test of the following hypothesis from the article “Gold’s Bluff – Is a 30 Percent Drop Next?” [no longer available]: “Ironically, gold is more than just a hedge against market turmoil. Gold is actually one of the most accurate indicators of the stock market’s long-term direction. The Dow Jones measured in gold is a forward looking indicator.” To test this assertion, we examine relationships between the spot price of gold and the level of the Dow Jones Industrial Average (DJIA). Using monthly data for the spot price of gold in dollars per ounce and DJIA over the period January 1971 through October 2024, we find that: Keep Reading