Double Bottom Pattern - A Systematic Approach

The identification of patterns in time series of stock prices is a fundamental aspect of financial market analysis. These patterns can be recognized and utilized through various approaches, including traditional technical analysis and advanced quantitative methods.

Chart analysis, a central subcategory of technical analysis, has traditionally relied on the subjective interpretation of analysts. This practice involved identifying visual patterns in the price movements of financial instruments to predict future market behavior. Historically, this analysis was heavily dependent on the analyst's experience and intuition, which could lead to variability in interpretations.

However, advancements in computer technology have brought about a significant transformation. Increasing computational power and the development of sophisticated algorithms have enabled the objective detection and quantitative evaluation of even complex chart patterns. This technological evolution has paved the way for more precise and systematic analysis of chart patterns.

Modern software and specialized applications employ algorithms capable of identifying recurring patterns and shapes in historical price data, free from subjective human biases. For example, algorithms can be used to detect specific patterns such as Double Bottoms, support and resistance lines, or trend channels. These patterns are then quantitatively assessed by calculating probabilities for future price trends based on historical data analysis and statistical models.

This objective approach not only enhances the accuracy of technical analysis but also broadens its applicability. Investors and traders can now make data-driven decisions based on solid statistical foundations rather than personal interpretation. This development marks a significant step toward a more rational and scientific approach to financial markets, significantly improving the reliability and effectiveness of technical analysis.

APPROACH AND OBJECTIVES

This project implements a specific approach to the automated detection of chart formations to conduct comprehensive backtests with detailed statistical analysis. One primary goal is to quantitatively evaluate the impact of various factors—particularly the height and duration—of a Double Bottom formation on trading performance.

To optimize the detection of chart patterns, this project utilizes the ZigZag indicator applied to price data. Unlike the usual percentage-based configuration, the ZigZag indicator here is determined based on twice the Average True Range (ATR), enabling a more dynamic adjustment to market volatility. A key criterion for updating the ZigZag indicator is that the current candle must represent the lowest price of the last eight trading days. This condition is checked daily to ensure the pattern accurately reflects significant turning points in the price trend.

Following the initialization of the ZigZag indicator, a window is defined for further analysis, covering a range of ±0.5 times the Average True Range (ATR) over 50 days. This window is overlaid on the chart to verify whether an additional low exists within this range. The first low identified within the window is considered the first low of the Double Bottom formation. Furthermore, it must be ensured that no other low occurred within at least 10 trading days before this identified low to confirm the validity of the formation

FIGURE 1: Detection of the Double Bottom

It should be noted that for more detailed analyses beyond the scope of this section, additional factors can be considered. These include, for example, the strength of the preceding downtrend before the first low, changes in trading volume during the formation, and possible divergences with the RSI14 indicator. These elements can provide additional insights into the robustness and reliability of the identified Double Bottom formation.

FORMATION PARAMETERS:

a) Identification of a current low (2nd low of the Double Bottom)

b) New low within the last 8 trading days

c) ±0.5 ATR distance within the last 50 trading days

d) No lower low 20 trading days prior to the 1st low

e) A maximum of 20 trading days after the 2nd low is allowed for the breakout to occur

Variants of the Double Bottom Formation

When trading these formations, two different entry strategies can be applied, each with its own advantages and risks:

1. Entry at the second low - Reversal:  This approach aims to enter at the second low of the Double Bottom formation, which is interpreted as a potential turning point in the price trend. The assumption here is that this point could initiate a renewed upward movement, allowing the trader to benefit early from the reversal.

FIGURE 2: Reversal Entry

2. Entry upon breakout above the neckline - Breakout: The breakout approach, on the other hand, waits for the price to break through the so-called neckline—a resistance line that represents the high point between the two lows.

FIGURE 3: Breakout Entry

In the breakout scenario, two additional variants can be considered for entry:

3. Breakout above a trendline: In addition to breaking the neckline, a breakout can also occur above a trendline. This trendline is defined by connecting the high point of the Double Bottom with the last high prior to the Double Bottom, selecting the one with the steepest slope. To minimize risk, the stop-loss (SL) can be set at the low of the Double Bottom, limiting potential losses in case the price falls back.

FIGURE 4: Trendline as a Signal Indicator

4. Double Bottom in an uptrend: Although a Double Bottom typically occurs at the end of a downtrend, the algorithm presented here can also identify Double Bottoms within an uptrend. In such a context, the formation is often considered a type of flag, signaling a continuation of the existing uptrend. This provides an additional variant for analysis and trading, as it confirms the potential strength and durability of the uptrend.

FIGURE 5: Double Bottom in an Uptrend

The Backtest and Statistics of the Double Bottom Formation

The Double Bottom is considered a significant signal for entering a trade. The position is opened at the beginning of the next trading day. For risk management, the stop-loss is set at the low of the Double Bottom, which is at least three times the average True Range (ATR) of the last 21 trading days.

The target for profit-taking (take-profit) is determined based on the risk-reward ratio (RRR), setting the potential profit at twice the possible loss. This strategic take-profit setup aims to ensure efficient profit realization while controlling risk.

Finally, a time limit for the trade is established: If neither the stop-loss nor the take-profit is reached within 60 trading days, the position is closed. This rule helps prevent capital from being tied up for too long in uncertain or unproductive positions and increases capital turnover.

TRADE PARAMETERS:

·     SL (Stop Loss): Low of the Double Bottom Formation

·     TP (Take Profit): 2 * RRR (Risk-Reward Ratio)

·     Time Stop: 60 trading days

·     Test Universe: S&P 500 stocks, historical composition

·     Time Period: January 1, 1996, to November 1, 2023

FIGURE 6: Entry and Exit, SL and TP

The signal for this backtest is the Double Bottom formation in a downtrend, with a breakout above the neckline.

For the backtest, the stock universe of the S&P 500 was used, based on historical data from its composition between January 1, 1996, and December 31, 2023. Over this period, the system generated a total of 3,148 trading signals.

The system’s hit rate, i.e., the proportion of profitable trades, was 51.2%, which is a solid value considering the risk-reward ratio (RRR) of 2.

The system's profit factor was 1.41. This value, representing the ratio of gross profits to gross losses, indicates that the trading system was overall profitable. Over the entire test period, the system achieved a performance gain of 106%. Despite this positive result, it is worth noting that the maximum drawdown—the largest observed loss from the last peak to the next trough—was 26%.

Challenges in Signal Distribution

Trading signals present challenges due to irregular signal distributions. At times, up to 82 signals may occur simultaneously, whereas at other times, signal frequency drops drastically. This results in fluctuations in capital requirements, as more capital must be reserved for simultaneous trades during periods of high signal activity.

To ensure effective money management and achieve a more balanced capital allocation, signal filtering is essential. This filtering regulates the number of simultaneously open positions and prevents portfolio overload.

Approach to Equity Curve Calculation and Risk Management

For the calculation of equity curves and risk control, a differentiated approach is applied to the various trading strategies:

Breakout Variants:

·     The average stop-loss (SL) is approximately 12%.

·     To limit the risk of a single trade to around 1% of total capital, the number of simultaneous open positions is capped at a maximum of 8.

Reversal Variants:

·     For these variants, the average SL is approximately 5%.

·     Up to 50 simultaneous positions are allowed.

If more signals are generated on a trading day than the number of open positions allowed, a selection process is applied. Priority is given to those signals with relative strength over 60 days and weaker evaluations. This selection criterion helps manage overall risk and focuses on potentially weaker signals that require less capital deployment.

This structured approach enables disciplined and risk-conscious portfolio management, which is essential for achieving long-term successful trading results and effectively allocating capital.

FIGURE 7: Equity Curves, 4 Variants

In Figure 7, the indexed equity curves (base 100) of the four trading variants (downtrend, trendline, uptrend, and reversal) are shown. The initial draft of the trading setup demonstrates promising results, especially in the breakout variants, which achieve high profits but also exhibit significant drawdowns. The reversal variant, while generating the most signals, performs worse compared to the breakout variants. Adjusting the strategy parameters or applying stricter signal filtering could improve the effectiveness and risk management of the setup. The performance and risk, measured by the maximum drawdown, are illustrated in Figure 8.

FIGURE 8: Return-Risk Diagram, 4 Variants

It can be observed that the trendline variant yields the highest profit, while the uptrend variant offers the best profit-to-maximum drawdown ratio.

Type	Downtrend	Trendline	Uptrend	Reversal (CRV=3)
Number of Signals	1086	1166	970	12616
Profit Factor	1.392	1.46	1.51	1.18
Hit Rate [%]	50.7	51.1	54.1	33.6
Profit [%]	639	957	547	127
Max Drawdown [%]	63	61.5	32	71
Max Simulaneous Positions	8	8	8	50
Average SL	12.4	9.7	11.8	4.8

TABLE 1: Overview of Backtest Results with Different Trading Setups

By filtering the signals, the number of trades can be effectively reduced while simultaneously improving the quality and reliability of the signals. This approach enables a more targeted selection of trading opportunities that promise a higher probability of success. The following chapter details how the signal quality is optimized, introducing various filtering techniques and criteria that contribute to further enhancing the performance of the trading system.

Optimization of the Double Bottom Formation

So far, the analysis of formations has primarily focused on price movements, while trading volume has been overlooked. Volume plays a crucial role in confirming reversal points and breakouts and can significantly contribute to optimizing the performance of chart formations. The underlying theory suggests that volume should confirm a trend. Accordingly, a healthy, rising trend typically exhibits increasing volume.

To determine whether volume increases up to the breakout, a regression line is calculated over the volume of the last five days. This analysis helps identify whether an upward trend in volume is present, providing additional confirmation of the trend's strength.

As a second condition for robust signal confirmation, it is required that the average volume during the breakout phase must be higher than the volume during the downtrend of the Double Bottom. This criterion supports the assumption that a strong breakout is accompanied by increased interest and heightened market activity, increasing the likelihood of a successful trade. By integrating these volume-based conditions, the reliability and performance of formations can be significantly improved.

FIGURE 9: Optimization of the Double Bottom in a Downtrend with two Volume Rules

By introducing a simple rule for volume, the performance of the trading setup can be approximately doubled, while the drawdown is nearly halved. This optimization demonstrates how effectively the integration of volume data into the trading strategy can enhance profitability while reducing risk.

FIGURE 10: Equity Curves for Downtrend, With/Without Optimization

FIGURE 11: Return-Risk Diagram, 4 Variants With/Without Optimization

Type	Downtrend	Downtrend Optimized
Number of Signals	1086	793
Profit Factor	1.392	1.7
Hit Rate [%]	50.7	54.3
Profit [%]	639	1413
Max Drawdown [%]	63	37
Max Simultaneous Positions	8	8

TABLE 2: Results of Optimizing the Downtrend Variant Using Volume

Summary

The development of an algorithm for pattern recognition has proven to be an effective tool for successfully identifying Double Bottoms. A key advantage of algorithmic pattern recognition lies in its ability to systematically capture and statistically evaluate all relevant conditions at the time of entry. This methodical data recording enables a deeper understanding of the circumstances under which certain formations yield particularly favorable trading results. Numerous historical examples highlight the effectiveness of this approach.

The combination of objective algorithmic pattern recognition and dynamic signal optimization offers promising opportunities for the further development and application of this trading approach. This integrated method ensures that the trading system remains stable and profitable even in changing market environments, thereby opening up new perspectives for future trading strategies.

Double Bottom Live Signals:

https://www.iqquant.com/iqqpatternrecognition