Technical Trading

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

Can investors reliably exploit the somewhat opaquely presented strategy summarized in “Using Wilder Volatility Stops to Time the U.S. Stock Market”, which employs Welles Wilder’s Average True Range (ATR) volatility metric to generate buy and sell signals for broad U.S. stock market indexes? To investigate, we each trading day for the SPDR S&P 500 ETF Trust (SPY):

1. Compute true range as the greatest of: (a) daily high minus low; (b) absolute value of daily high minus previous close; and, (c) absolute value of daily low minus previous close.
2. Calculate ATR as the simple average of the last five true ranges (including the current one).
3. Generate a Wilder Volatility Stop (WVS) by multiplying ATR by a risk factor of 2.5.
4. When out of SPY, buy when it closes above a dynamic trendline defined by a trend minimum plus current WVS (breakout). When in SPY, sell when it closes below a dynamic trendline defined by a trend maximum minus current WVS (breakdown).

We perform the above calculations using raw (not adjusted for dividends) daily SPY prices, but use dividend-adjusted prices to calculate returns. We assume any breakout/breakdown signal and associated SPY-cash switch occurs at the same close. We initially ignore SPY-cash switching frictions, but then test outcome sensitivity to different levels of frictions. We ignore return on cash due to frequency of switching. We further test outcome sensitivity to parameter choices and to an alternative definition of ATR. We use buy-and-hold SPY as a benchmark. Using daily raw and dividend-adjusted prices for SPY during January 1993 (inception) through most of October 2024, we find that: Keep Reading

Can pattern recognition software, combined with news sentiment, identify profitable short-term stock trades? In their July 2024 paper entitled “Technical Patterns and News Sentiment in Stock Markets”, Markus Leippold, Qian Wang and Min Yang test the ability of pattern recognition software (convolutional neural network) to find profitable technical reversal patterns within U.S. and Chinese stock candlestick charts. They consider four pairs of technical patterns: double tops and bottoms; head and shoulders and inverted head and shoulders; broadening tops and bottoms; and, triangle tops and bottoms. They use Bollinger bands to find local maximums and minimums, with the standard deviation multiplier set at 1.1 based on parameter tuning. They augment pattern recognition with news sentiment from Refinitiv for U.S. stocks since 2003 and from Tonglian for Chinese stocks since 2014 during the 10 trading days around each pattern. They identify combined tops as double tops, head and shoulders, broadening tops or triangle tops coupled with negative news and combined bottoms as double bottoms, inverted head and shoulders, broadening bottoms or triangle bottoms coupled with positive news. They first consider each technical pattern as an independent event and measure abnormal returns for holding intervals of 1, 5, 10, 21 and 42 days after a signal. They then examine the performance of a portfolio of events for a 1-day holding interval. They use U.S. stock data from 1992 to 1999, enhanced via two data augmentation strategies, for pattern recognition software training and validation. They then apply the trained software to U.S. stock data from 2000 through 2021 and Chinese stock data from 2005 through 2021. Combined pattern-sentiment test periods are shorter based on availability of sentiment data. Using price series for all U.S. common stocks and Chinese A-shares and news sentiment data as described through 2021, they find that: Keep Reading

Can investors manage cryptocurrency volatility risk with simple trend following rules? In their August 2024 paper entitled “Trend-following Strategies for Crypto Investors”, Trinh Hue Le and Ummul Ruthbah test simple trend following rules on Bitcoin (BTC), Ether (ETH) and the S&P Cryptocurrency LargeCap-Ex. MegaCap Index. Specifically, they hold the cryptocurrency (cash) when its price is above (at of below) its prior-day simple moving average (SMA) of 20, 65, 150 or 200 days. To assess net profitability, they consider trading frictions of 0.1%, 0.25% or 0.5% of amount traded. They further measure correlations between the movements of cryptocurrencies and those of the Nasdaq 100 Index. Using daily prices for the S&P BTC, ETH and Cryptocurrency LargeCap-Ex. MegaCap indexes starting January 2016, January 2017 and January 2019, respectively, all through January 2023, and contemporaneous daily levels of the NASDAQ 100 Index, they find that:

Keep Reading