Bollinger Band Revisited — How Would an AI Trade It?
What is a Bollinger Band?
Bollinger Bands are a well-established tool in technical analysis, developed by John Bollinger in the 1980s. These bands consist of three lines: a middle line, which is a simple moving average (SMA) typically calculated over 20 periods; and the upper and lower bands, set at two standard deviations from the middle SMA. This configuration is meant to help traders assess market volatility and potential conditions where assets are overbought or oversold. Specifically, if the price hits the upper band, the asset might be considered overbought, and if it hits the lower band, oversold. This information is considered helpful for traders looking to determine entry and exit points to capitalize on potential price reversals.
A Classic Bollinger Band Strategy
Let’s examine how such a strategy would have performed over the past year for the NVIDIA stock, using end of day data and a five-day holding period following the last signal. As a tool, we will use QUINETICS, allowing us to create algorithmic trading strategies with one click.
After selecting NVIDIA, we can select the “Classic” model type and “Bollinger” as an indicator.
Now let’s look at the results of the strategy. The red dots are sell signals (we short the stock), generated when the price climbed above the upper band. The green dots are buy signals generated when the stock price fell below the lower band.
And the backtest of this strategy:
Initially, the results were promising; the strategy yielded about a 25% return by the end of December 2023. However, in 2024, relying solely on Bollinger Bands would have eradicated earlier gains, underscoring a significant drawback of using only one reversal indicator. Specifically, if the stock price continues to climb, traders miss out on the momentum or even incur losses if they opted to short sell.
An AI-Enhanced Bollinger Band Strategy
Now, let’s switch to AI models to interpret the indicators.
QUINETICS trains AI models that are eventually deployed to interpret new data to predict future returns. Users can create their strategies based on these machine learning models. Let’s have a look how the AI would have traded the information for the Bollinger Band indicator in the past year:
When volatility increased — evidenced by the bands diverging significantly from the moving average — the AI stopped issuing buy signals (the green dots in the chart above). This adaptation helped the AI avoid a market downturn in recent weeks. The results?
The backtest shows a 231% return over the past year, with a Sharpe Ratio close to 3. Note that backtesting results are not predictive of future performance.
Conclusion
While traditional Bollinger Bands may offer insights into market conditions, the AI-enhanced approach in our example fared better. Moving forward, traders might consider combining AI with traditional indicators to enhance decision-making processes.
If you want to see more content like this, follow us on social media:
📹Subscribe to our YouTube channel
🐦Follow us on X/Twitter
🏢Follow us on LinkedIn
🔔Register for free on QUINETICS
Cheers!
Fabian
Visit us at DataDrivenInvestor.com
Subscribe to DDIntel here.
Featured Article:
Join our creator ecosystem here.
DDI Official Telegram Channel: https://t.me/+tafUp6ecEys4YjQ1
