Short-term Crossover Tracking with BTC/USDT

Identifying trends using Moving Average crossovers on real-time BTC/USDT prices | Trend-following Trading Strategy

About The Project

• This project aims to gain insights into market dynamics by using Simple Moving Average (SMA) and Exponential Moving Average (EMA) over different timeframes as technical indicators to predict future price trends
• By comparing the Moving Averages and the current mid-price, instances where the price rises above or falls below the Moving Averages - crossovers, might signal potential momentum or market sentiment
• Instances where the short-term moving average crosses the long-term moving average like The Golden Cross or The Death Cross also signals a bullish/bearish breakout pattern
• From the partial order book, we are also able to access the liquidity of the market and overall ease for traders to enter and exit positions by calculating the bid-ask spread

Real-time BTC/USDT price data from Binance

• KLine/Candlestick Stream
• Partial Book Depth Stream

Technical Implementation

• Real-time, low latency WebSocket communication is used to consume highly-frequent market data from Binance, whilst minimising overhead when maintaining the persistent connection

• Data is instaneously processed and written into csv files for further post-analysis

• The closing price from the 1min KLine stream data is used to calculate the 5mins Moving Averages while the closing price from the 5min KLine stream data is used to calculate the 25mins Moving Averages

• The mid price is calculated by taking the average of the best bid and the best ask price from the Partial Book

• Crossover Indicator

  • OVER: Mid price rises above the SMA
  • UNDER: Mid price falls below the SMA
  • NEUTRAL: Mid price is equal to the SMA

What is Simple Moving Average (SMA)

• It is a calculation that represents the average over a set of prices within a certain timeframe
• It can act as a technical indicator that can aid in determining if an asset price will continue in its current trend and direction or not
• It smoothens out short-term price fluctuations and highlights the general direction of the trend

Learn more

What is Exponential Moving Average (SMA)

• It is another type of moving average calculation
• Compared to SMA, EMA gives more weight to recent prices, making it more responsive to recent market changes
• This might be more suitable for traders focused on short-term movements as EMA reacts more quickly to changes in price

Learn more

Built With

• Python
• Tornado

Getting Started

1. Install dependencies required pip install -r requirements.txt
2. Run the application

   cd app
   python main.py
   

3. WebSocket handlers

   ws://127.0.0.1:8888/binance-short-kline
   ws://127.0.0.1:8888/binance-long-kline
   ws://127.0.0.1:8888/binance-partial-book
   

Results

Timestamp	Mid Price	5mins SMA	5mins EMA	25mins SMA	25mins EMA
2024-01-17 08:09:34	42592.9	42592.90400000001	42597.01009266221	42663.566000000006	42665.98251851853
2024-01-17 08:09:35	42594.295	42592.90400000001	42597.01009266221	42663.566000000006	42665.98251851853

Instance of mid price rising above 5mins SMA, potentially signal an upward momentum as price rises above the "historical average"

Comparing 25mins Simple Moving Average and 25mins Exponential Moving Average

• EMA reacts faster to price changes than SMA

Comparing 5mins and 25mins Simple Moving Average

• Shorter timeframes can react more quickly to price changes
• Longer timeframes can provide a smoother trend indication but may lag in responding to rapid price movements

Backtesting

• Using historical kline data from 01-01-2024, we applied the 5mins and 25mins SMA Crossover to derive trading actions
• If the 5mins SMA rose above the 25mins SMA, we would perform a buy, and conversely if the 5mins SMA dipped below the 25mins SMA, we would perform a sell
• Terminal return: -0.927%
  • This could have been impacted from highly volatile short-term fluctuations where short-term trends do not reflect sufficient momentum to see profit

Green arrows indicate a buying action and Red arrows indicate a selling action

Modified Moving Average Intervals

• Terminal return: +7.083%

• Using a longer timeframe and moving average of 3 days and 10 days, we were able to see much better returns
• While this would be an unfair comparison to make with the initial test due to the vast difference in time horizon, it still shows that Trend-Following can be a valid trading strategy given the right parameters

Improvements

Likelihood of false signals

• False signals can be more prevalent in choppioer markets and additional indicators and filters should be used to reduce the liklihood of acting on false signals