In the rapidly expanding universe of cryptocurrencies, Binance's token, BNB, stands out as one of the most influential and widespread digital coins. Originally launched in 2017 by global exchange Binance, BNB was conceived as a utility token that offers users discounts on trading fees. Since then, its application has expanded far beyond just being used on Binance, becoming the cornerstone for a wide range of financial operations and decentralized applications on the BNB chain.
With a market capitalization that regularly ranks among the top five cryptocurrencies in the world, BNB has established itself not only as a heavyweight currency, but also as a barometer of the financial health of the entire crypto sector. This prominent position compared to other digital currencies such as Bitcoin and Ethereum, which dominate the market by capitalization, puts BNB in a unique position to capitalize on market fluctuations and new trends in cryptocurrency.
In this article, we will explore how the distinctive features of BNB can be used to build a robust algorithmic trading strategy based on a reversal approach using Bollinger Bands.
Bollinger bands
This indicator takes its name directly from its inventor, John Bollinger, who analyzed the behavior of prices as they moved away from or closer to their moving average. Bollinger chose to include two bands, calculated as the standard deviation of the simple price average. Bollinger bands are therefore made up of 3 elements and are calculated using the following mathematical functions:
- Upper band = average price over the last N periods plus 2 standard deviations;
- Medium price = average price over the last N periods (20 is the recommended starting number);
- Lower band = average price over the last N periods minus 2 standard deviations.
Reverse trading strategy (Bollinger Bands) applied to BNB
The strategy that will be adopted is an automatic system with the logic of “reversion to the mean” that uses Bollinger Bands as the turning point of the market. When prices reach the upper band, we will sell, while at the lower band we will buy.
The considered session usually starts at 00:00 GMT and ends at 23:59 GMT to coincide with the solar day, as cryptocurrencies are quoted 24 hours a day.
Assuming you operate with $10,000 per trade, the trade will be closed when the profit target of $3,000 is reached, which is the first value. It will also be very useful to use a fixed stop loss in the beginning, which we assume to be $1,000. The stop loss should somehow protect the invested capital from trades with very high losses.
By applying this strategy to the BNB/USDT pair on a 15-minute time frame, it is possible to see how this “operating engine” would perform from 2021 to today. Previous data is not taken into account, since from 2021 BNB began to establish itself among the main cryptocurrencies in the world. This phenomenon is also due to the strong overall growth of the crypto world and the DeFi sector in particular, with a significant increase in total value locked (TVL) in various platforms based on the BNB Chain. BNB was used as the main token for transactions and management on many platforms, which allowed an increase in demand and subsequently its price, which in 2020 increased to more than 15 times the average value.
Results of the reversal strategy (Bollinger Bands) applied to BNB
Figures 2, 3 and 4 show the metrics obtained from the mean reversion strategy based on Bollinger Bands: the results are encouraging with the rising equity line. That's definitely a good starting point.
On a closer analysis of the results, it is possible to notice that the average trade is around 41 USD, which, compared to the amount of one operation (10,000 USD), corresponds to 0.41%, a value that does not guarantee the coverage of operating costs. .
Optimization of the reverse trading strategy (Bollinger Bands) applied to BNB
Trying to define a different operating window may lead to a better result. That is why we proceed to optimize the start time of the operation and its duration (expressed in the number of 15-minute cycles). The results of the optimization can be seen in Figure 5. When operating from 00:00 until the end of the next 60 bars or until 15:00, the values improve significantly: the total profit of the system rises to $222,450 with more than 30% fewer operations ( 2,875), in as a result, the average trade rises to $77.
The results are definitely better, but still indicate a harsh strategy with a larger average trade, but still not too high, and a fairly high drawdown compared to net profit (Net Profit/Max DD = 3.6).
To proceed further, we could try to optimize the originally assumed stop loss and take profit values. In fact, the first stop loss optimization confirms that $1,000, which is 10% of the position value, achieves interesting results. Moving on, Figure 6 shows that changing the take profit in $500 increments produces interesting and rather stable results around $5,000 to $6,000. Increasing these values improves the metrics, but not in a very significant way, so one could choose to use $6,000, for example.
Adding patterns to the reversal strategy (Bollinger Bands) applied to BNB
Since the strategy still executes many trades (about 2,800 in 4 years), there is probably still room for further filtering of trades. To do this, you can use some price formulas that can identify the best conditions for performing operations and filter out those with a lower probability of success.
For this purpose, we will use a proprietary list that contains many different price combinations, which will help us understand in which situations BNB seems to respond better to the input logic of this system.
By analyzing different combinations of patterns, for example, it was found that by inhibiting operability when the patterns “PtnLN=19” occur for the long side and “PtnSN=11” for the short side, a good trade-off is achieved between the main benchmarks (Net Profit, Average Trade, Max Intraday Drawdown). With the PtnLN=19 pattern, we avoid going long if the close of the last session candle was greater than 0.5% of the close 2 sessions ago. With PtnSN=11, it will instead prevent short entry simply if the previous session confirmed an ongoing downtrend (indicating both a lower low and a lower high).
It is noted how this combination of filters can actually increase the average trade ($240) and the net profit rises above $400,000. Even the drawdown goes down to -$39,702. A good improvement, also evident in the more regular shape of the equity line (Figure 8), although it seems to have lost some luster in the latter part of 2023. Only time will confirm the correctness of the choices made during the optimization phase.
Conclusions on the strategy
The reversal strategy with Bollinger Bands has proven to be quite interesting on the BNB/USDT pair, although it would need further refinement to be ready to work live on the market. This particular cryptocurrency, although established, is still relatively young and presents many opportunities for traders who want to try different approaches to the market. As always, it is left to the reader to experiment and develop this idea to its fullest.
Until next time and happy trading!
Andrea Unger
