A Trading System Based on Polynomial Regression Models

Subscribe to newsletter

Linear regression is a widely used prediction technique in finance. Linear regression can estimate the relationship between a dependent variable (such as stock price) and one or more independent variables (like market indices or economic indicators). This approach is particularly useful in predicting trends, asset prices, and risk factors. However, while it is effective for identifying linear relationships, it may be limited when handling complex or non-linear data patterns, which are also common in financial markets.

Reference [1] proposed using polynomial regression models to predict stock price movements. Specifically, it applied two-, three-, and four-degree polynomial regression models to forecast stocks’ daily movements, and used two standard deviations from the regression model to generate trading signals. The trading strategy works as follows,

  • When the stock’s closing price crosses above the upper band, the system purchases the stock at the opening price on the next trading day, thus initiating a long position.
  • When the closing price touches the lower band, the system sells the stock at the opening price the following day, closing the long position.

The author pointed out,

Subscribe to newsletter https://harbourfrontquant.beehiiv.com/subscribe Newsletter Covering Trading Strategies, Risk Management, Financial Derivatives, Career Perspectives, and More

In this research, we explored whether a trading system based on polynomial moving regression bands (MRBs) could be effectively applied to Nasdaq100 stocks to trade profitably. We designed three fully automated trading systems using polynomial MRBs of degrees two to four and tested their performance on daily stock data from early 2017 to the end of March 2024. To limit the overfitting problem, we employed several strategies: we used lower-degree polynomials to avoid unnecessary complexity, implemented early stopping during model training to prevent the models from becoming too specialized to the training data, and conducted random searches to find the optimal balance between model complexity and performance. Our results indicated that all three models achieved, on average, a positive net profit (NP), with a percent of profitable trades (PP) exceeding 50% and a profit factor (PF) greater than 1, demonstrating that the behavior of Nasdaq100 stocks can be effectively modeled using polynomial regression without overfitting. Among the models, Model 3, which employed a fourth-degree polynomial MRB system, performed the best, achieving the highest average NP of USD 162.73 and exhibiting a superior risk profile with the lowest average loss (USD 95.52), highest minimum PP (41.51%), and highest minimum PF (0.55). These findings suggest that carefully designed polynomial MRB-based trading systems, which mitigate overfitting through model selection and validation techniques, can successfully capture stock price trends and be used to trade profitably in fully automated systems.

In short, using higher-order regression models proved effective.

While this idea has its merits, it’s worth noting that as the degree of the polynomial increases, the risk of overfitting also rises. To mitigate this risk, rigorous out-of-sample testing is essential. Although the author claimed that out-of-sample validation was performed, the description of this procedure was too brief. More details on the out-of-sample validation would be desirable.

Let us know what you think in the comments below or in the discussion forum.

References

[1] Gil Cohen, Polynomial Moving Regression Band Stocks Trading System, Risks 12: 166, 2024

Subscribe to newsletter https://harbourfrontquant.beehiiv.com/subscribe Newsletter Covering Trading Strategies, Risk Management, Financial Derivatives, Career Perspectives, and More

Further questions

What's your question? Ask it in the discussion forum

Have an answer to the questions below? Post it here or in the forum

LATEST NEWSWhich industries will feel the pinch of Trump’s tariffs the hardest?
Which industries will feel the pinch of Trump’s tariffs the hardest?

OTTAWA — U.S. President Donald Trump signed an order Saturday to impose tariffs of 10 per cent on Canadian energy and 25 per cent on everything else beginning on Feb. 4. Statistics Canada says the value of Canadian exports to the United States in 2023…

Stay up-to-date with the latest news - click here
LATEST NEWSTrump Tariffs Threaten to Roil Oil Market, Raise US Pump Prices
Trump Tariffs Threaten to Roil Oil Market, Raise US Pump Prices

US President Donald Trump’s tariffs on Canadian and Mexican imports threaten to disrupt North America’s tightly integrated oil market and push up gasoline prices for American motorists.

Stay up-to-date with the latest news - click here
LATEST NEWSTrump's calendar: When tariffs, RTO, buyouts, and a TikTok cut-off are set to go into effect
Trump's calendar: When tariffs, RTO, buyouts, and a TikTok cut-off are set to go into effect

Trump's deadlines for tariffs, federal worker buyouts and RTO, federal funding freeze, TikTok, and more are coming up. Here's a timeline of key dates.

Stay up-to-date with the latest news - click here
LATEST NEWSHere's how tariffs on Canada, China and Mexico may impact U.S. consumers
Here's how tariffs on Canada, China and Mexico may impact U.S. consumers

President Donald Trump signed orders on Saturday placing tariffs on Canada, China and Mexico. Here's how that may affect U.S. consumers.

Stay up-to-date with the latest news - click here
LATEST NEWSNew tariffs on Mexico, Canada and China announced by President Trump
New tariffs on Mexico, Canada and China announced by President Trump
Stay up-to-date with the latest news - click here

Leave a Reply