The RalliV1 Trading strategy that might boost your trading results (Patreon)

Introduction

Hi there and welcome back to this post where I dive deep into cryptocurrency trading strategies to help you become a more informed and effective trader. 

Today, I’m exploring a Freqtrade strategy known as RalliV1 of a coder called Rallipanos. All credits go to the author of the code, I’m just the guy who show his backtest results on the Youtube here.

Now, whether you’re new to trading bots or looking to refine your existing strategies, this session  has some with valuable insights of which I hope you can use for refining your own trading algorithms!

The strategy

Let’s discuss the strategy code and the intentions of it.

After we come accross the mandatory settings for use with the Freqtrade bot, like the loading of libraries, defining hyperoptable buy and sell space parameters. The ROI table and stop loss setting, and some more protection and buy/sell settings we come to the timeframe.

Now, the RalliV1 is a dynamic strategy, initially designed for the 5-minute timeframe, with the help of a 1 hour informative timeframe so it seems.

The custom stop loss is designed to cut losses quickly if a trade isn't performing as expected. Specifically, if a trade's profit is less than 0.1% after 140 minutes from the trade's opening, the strategy forces a stop loss at a 0.5% loss to prevent further downside risk. This method helps manage losses proactively, ensuring that unsuccessful trades are exited early to preserve capital.

It leverages multiple indicators to make precise entry and exit decisions. It also focuses on combining short-term price movements with various technical indicators to optimize buy and sell signals.

Firstly, the strategy employs Exponential Moving Averages (EMAs) and Simple Moving Averages (SMAs) to determine the market trend and potential reversal points. It uses these in conjunction with the Hull Moving Average for additional trend confirmation.

Next, the Relative Strength Index or RSI is utilized in two forms—fast and slow—to gauge the momentum and identify overbought or oversold conditions.

It also incorporates an Elliott Wave Oscillator (EWO), calculated from two different EMAs, to help in recognizing price momentum and the end of certain price patterns.

Some of these indicator settings can be adjusted to fit the pairs behaviour well by using hyperoptimization. However using this will also add bias, called curve fitting, to the performance results of the backtest since it tries to fit the past extremely well. Creating the danger that these settings will not match possible changes in character of the pairs you will use. So please use this with caution.

Buy Signals

Let’s now talk about the buy and sell signals here.

For buy signals, RalliV1 looks for conditions where the price is below a specific EMA and SMA, indicating a potential undervalued state. It also checks that the fast RSI is below a threshold to ensure it's not entering during overbought conditions.

The strategy also employs an Elliott Wave Oscillator value to determine the right momentum before committing to a buy. This ensures the strategy buys into strength rather than weakness.

There are multiple buy conditions configured, each set with it’s specific buy logic. This means that the strategy can buy under certain set different circumstances. And can make the algo more agile to changing markets.

Sell Signal

The same counts for the sell logic.

Sell signals are generated when the price moves above certain moving average thresholds and the fast RSI indicates a shift from slow RSI, suggesting a potential overbought condition and a good point to take profits.

And you can see that there are two sell conditions configured. Each with their own tiny differences. 

Backtest results

Ofcourse I have tested this algo with my standarized way of testing to see how the hypothetical performance looks and how that performance is compared to other trading strategies I tested earlier.

Wins, Draws, and Losses:

The cumulative wins, draws, and losses chart reveals that the number of winning trades does grow over time, which is a positive sign. However, the cumulative profit line flattens out significantly towards the latter part of the chart, suggesting that the strategy may not be capitalizing on its winning trades effectively or that it is giving back a lot of gains during the losing trades.

Drawdown Analysis:

The drawdown analysis plot exhibits periods of relatively high drawdown, with the most severe drawdowns occurring around the bull market peak of 2021. Notably, while the average drawdown hovers around 3.24%, the spikes well beyond this average indicate that the strategy may not be consistently managing risk effectively. This could be a red flag for potential users of the strategy, as large drawdowns can significantly impact the trading account, especially if these occur in rapid succession.

Profit/Loss Over Time:

The mean profit/loss plot underscores the strategy's inconsistency, with profits not showing a stable upward trend. While there are periods of profitability, they are interspersed with significant losses. The plot shows that while the strategy is capable of capturing gains, it does not do so reliably over time. This could suggest that the strategy is sensitive to market conditions and might require additional safeguards or adjustments to improve its consistency.

Winrate and Profit Distribution:

The winrate and profit distribution box plots present a clearer picture of the strategy's performance. The winrate appears to be favorable; however, a good winrate does not necessarily translate to high profits, as is evident from the profit distribution plot. The profit distribution exhibits several outliers which indicate substantial losses, signifying that the strategy could be prone to large, infrequent negative trades that could erase a significant portion of the gains.

Performance Indicator Comparison:

It shows that while RalliV1 may not top the charts in every category, it holds its own with a respectable win rate, a clear indication that the strategy has moments where it shines.

When we consider metrics like the Calmar, Sortino, and Sharpe ratios, which are a bit on the lower side, it's important to remember these are just one piece of the puzzle. The real-world application of a strategy can sometimes diverge from backtest metrics.

Moreover, the strategy's overall score places it within the same ballpark as other well-performing strategies. This suggests that with some refinement and active management, RalliV1 has the potential to climb the ranks even more. It's not at the bottom of the pack, and in the world of trading, strategies often require a blend of solid mechanics and skilled execution.

RalliV1's profit factor indicates that the strategy has been profitable over the backtesting period. Combined with its win rate, this could suggest that with further tuning, RalliV1 might increase its efficiency.

So, while RalliV1 may not be the current leader in the strategy league, it's certainly not without merit. With the right tweaks and a careful approach to risk management, there's potential to unlock greater performance and reliability.

Strategy League & conclusion

So with these results on the 15 minute timeframe I can argue that I have found another trading algorithm that could be well suited for use in an automatic trading bot. It’s score currently put this trading strategy on a proud third spot of the current strategy league.

But again be aware that this is just a theoretical excersise and real world trading should only occur if you do your own rigourous testing on your own setup and also after a reasonable period of forward testing. Only after good results in these situation you can carefully think about putting real world money on the line with manual or real world trading.

So this is it for this video. I want to thank you for viewing. If you are testing or using this algo, then i’d really like to know if my findings are correct and you can confirm or debunk these results.

Furthermore I want to ask you to click the like button if you like the video.

Again thanks for watching, and I will see you in the next video.

Goodbye!