Backtesting OVER 200 Cryptos with Python in the current 📉 market [Part I]

Summary

In this video series, we will backtest over 200 cryptocurrencies in a downtrending market to evaluate a trading strategy. By analyzing data from a short 80-day period, we aim to identify the top five performers, which will undergo a 30-day validation period with the same strategy. We'll build a crypto price database using SQLite, implement a crossover long-only strategy, and explore data resampling techniques. Throughout, we emphasize that past performance does not guarantee future results, encouraging viewers to engage with the content while learning about effective cryptocurrency trading strategies.

Takeaways

• 📈 This video series will backtest over 200 cryptocurrencies in a downtrending market using a trading strategy.
• 🔍 After an initial evaluation of 50 days, the top five performing cryptocurrencies will be selected for a 30-day validation period.
• 📅 The look-back period is set at 80 days, starting from when all cryptocurrencies began to decline.
• 💾 Data will be stored in a SQLite database to manage the large volume of data efficiently.
• 📊 The series will utilize a simple crossover long-only trading strategy, with the potential to explore other strategies in the future.
• 🛠️ Key concepts will be covered, including creating a crypto price database and resampling data from one-minute to other time frames.
• 📉 The goal is to achieve a non-overfitting backtest and determine the profitability of the trading strategy over selected intervals.
• 📝 Previous knowledge from earlier videos will be referenced, with additional resources provided in the video description.
• 🚫 A disclaimer emphasizes that the strategies discussed are not investment advice and past performance does not guarantee future results.
• 🔗 The video includes practical examples, including data retrieval from the Binance API and resampling logic for open, high, low, and close data.

Q & A

What is the main objective of the video series?

The main objective is to backtest over 200 cryptocurrencies in a downtrending market using a trading strategy and evaluate the top performers.

Why is the lookback period set to 80 days?

The 80-day lookback period is chosen because it marks a significant point in time when all cryptocurrencies began to decline.

What kind of data is being used for analysis?

The analysis uses granular one-minute interval data for various cryptocurrencies, which is then aggregated into larger time frames.

What is the role of the SQLite database in this process?

The SQLite database is used to store the large amount of historical price data, as it is impractical to request all the data from an API repeatedly.

What trading strategy is primarily being applied in this series?

A crossover long-only strategy is being applied for the backtesting of cryptocurrencies.

How are the performance of cryptocurrencies evaluated?

After a 50-day evaluation period, the top five cryptocurrencies are selected based on their performance with the trading strategy.

What steps are taken to resample the price data?

To resample the price data, different aggregation methods are applied: the first value for open, the maximum for high, the minimum for low, and the last value for close.

What is the purpose of using the Binance API?

The Binance API is used to pull historical price data for cryptocurrencies, which is essential for the backtesting process.

What happens if the loop for importing data is interrupted?

If the loop is interrupted, a query can be run to check which symbols have already been imported into the database.

What disclaimer is provided regarding the content of the video series?

The disclaimer states that the concepts presented are not investment advice and that past trading strategies may not guarantee future returns.