Enhancing Trading Diversification With Bollinger Bands: A Mean-Reverting Strategy For Futures

To increase portfolio diversification, in this article, we will evaluate using another very well-known indicator in the trading landscape: Bollinger bands.

This indicator is named after its inventor, John Bollinger, who analyzed the behavior of prices when they moved away from or approached their moving averages. Bollinger added two bands calculated as the standard deviation of the simple average of prices.

With his famous bands, John Bollinger tried to find a way to confine prices within clearly defined price bands.

The two bands, the upper and the lower band, can become the “trigger” for entering a strategy: a departure from the bands signifies a breakout in volatility, so prices should continue their run in the direction taken, or reverse their predetermined course and complete a reversal if a departure from the bands is followed by a sudden re-entry into the bands.

The Bollinger bands consist of 3 elements and are calculated using the following mathematical functions:

  • UpperBand = average price of the last N periods plus two standard deviations;
  • MedianPrice = average price of the last N periods;
  • LowerBand = average price of the last N periods minus two standard deviations.

An image of the above indicator is shown in Figure 1.

fig_1.png

Figure 1: Bollinger Bands Indicator.

The strategy we will use is an automatic system with "mean-reverting" logic, i.e., we will exploit the Bollinger bands as the turning point of the market. Upon reaching the prices on the upper band, we will sell, while we will buy at the lower band.

Since this is a "mean-reverting" strategy, it is very useful to use a stop loss from the beginning, which can protect our capital from very high loss trades.

A stop loss, calculated with 3 times the average volatility of the last 5 bars, is introduced to balance the results. We will then test this strategy on a basket of futures to see how they would have performed from 2010 to 2022.

We will use a 30M (minute) time frame and test the following markets:

  • Crude Oil (CL)
  • S&P500 (ES)
  • DAX (FDAX)
  • Gold (GC)
  • Live Cattle (LC)
  • Soybeans (S)
  • US Treasury Note 30 Yrs (US)
  • Wheat (W)
  • British Pound (BP)
  • Bund (FGBL)
  • Copper (HG)
  • Heating Oil (HO)
  • Natural Gas (NG)
  • RBOB Gasoline (RB)
  • EuroFX (EC)
  • Feeder Cattle (FC)
  • Coffee (KC)

Figures 2, 3, and 4 show the metrics obtained with the reversal strategy based on the Bollinger bands. The results are encouraging.

The equity line is rising, which is undoubtedly a good starting point.

fig_2.png

Figure 2: Portfolio equity line produced by the Reversal BB strategy.

Moreover, there seems to be some consistency across the years, which indicates the quality of the information provided by the indicator.

Looking at the results for individual markets, we find that only 5 out of 17 markets suffered losses. As the analysis continues, it becomes apparent that specific markets seem to prefer  this kind of logic, namely Dax (FDAX), Gold (GC) and EuroFX (EC).

fig_3.png

Figure 3. Overview by individual market.

Note the total average trade, which reaches only $10. The value is so low because the strategy produces so many trades. This value is obviously not enough to make the system useful for live (in real) trading, as commission costs and slippage would completely erode it. The advantage is that there is still room to apply additional filters to the entries.

fig_4.png

Figure 4. Average trade of the portfolio.

Of course, these results could be the product of a single case. When optimizing indicators, it is always advisable to use time frames longer than 30 minutes (to avoid the market "noise," which would be too great if the bars were too fast) and not to go too far in optimizing the various inputs.

We want to see if limiting trading only to the days when a certain price pattern occurred can improve the average trade, which is currently the sore point of the system. So let us try to improve the average trade of the strategy without altering the indicator. For this purpose, we will use our own list of patterns (predefined set of filters) that includes different case histories to assess which specific situation is best for the portfolio.

Strictly speaking, a mean-reverting strategy works after the market has experienced days with wide movements in one direction or another. Therefore, we try to build a filter into our system that allows us to enter the market only when that specific condition is met. To select the condition, we start an optimization that determines which pattern produces better results than the others.

From the results in Figure 5, we can see that pattern 47 seems to be the best (in terms of profit). It represents a situation where the range (distance between the high and low) of the previous day is greater than the average of the last two days. In this way, we wait for an increase in the volatility/direction of the underlying asset during the last days to enter in counter trend, hoping that the market will reverse its course.

This pattern adds value and quality to the strategy. The total profit increases from $480,000 to $685,000, with a reasonable reduction in DrawDown. It increases from about 49,000 trades with an inactive filter (value 0) to about 41,000 trades with an active pattern filter. In short, 15% of the initial trades are cancelled, and the average trade rises to $16.

So, achieving a victory is impossible, because the value of the average trade is still too low for live trading, but this strategy can be a good starting point for further development.

fig_5.png

Figure 5. Pattern Optimization.

In summary, with such a varied and diversified portfolio, achieving a very high average trade would certainly be difficult. That would be a bit of an overreach of our strategy, which can hardly work well as is in all markets. However, with the help of some additional filters, Bollinger bands have proven that they can be an excellent arrow in the bow of any systematic trader.

Posterity will be the judge of that. I hope what you have seen has given you some interesting insights. My recommendation is to always be curious to find out if there are pitfalls hidden behind seemingly very positive results.

See you next time!

Happy trading,

Andrea Unger

Market News and Data brought to you by Benzinga APIs
Comments
Loading...
Posted In: FuturesMarketsTrading IdeasGeneralcontributorsJohn BollingerPortfolio Diversification
Benzinga simplifies the market for smarter investing

Trade confidently with insights and alerts from analyst ratings, free reports and breaking news that affects the stocks you care about.

Join Now: Free!