Bitcoin Multi-factor Trading Strategy

Author: ChaoZhang, Date: 2023-09-25 18:24:02
Tags:

Overview

This is a comprehensive trading strategy designed for 15-min timeframe trading of Bitcoin and other cryptocurrencies. It combines multiple indicators to generate buy and sell signals, including Triple Exponential Moving Average (TEMA), Average True Range (ATR), and Heikin-Ashi candles, together with risk management features like take profit and stop loss.

Strategy Logic

The strategy utilizes the following indicators:

  • Triple Exponential Moving Average (TEMA): Three TEMA lines of different lengths and sources, based on high, low and close prices respectively.

  • Average True Range (ATR): Custom ATR calculation with EMA smoothing to measure volatility.

  • Supertrend: Calculated using ATR and a multiplier to determine trend direction.

  • Simple Moving Average (SMA): Applied on the short TEMA line to smooth its values.

  • Heikin-Ashi Close: Used for additional trend confirmation.

Long entry signal is triggered when the short TEMA is above both long TEMA lines, Supertrend is bullish, short TEMA is above its SMA, and Heikin-Ashi close is higher than previous close.

Short entry signal is triggered when the opposite conditions are met.

Take profit and stop loss are set at 1% and 3% of entry price. Commission is also considered.

Advantage Analysis

  • Multiple factors improve accuracy Combining trend, volatility, pattern indicators can improve accuracy and avoid false signals.

  • Reasonable stop loss/take profit controls risk Well-set stop loss and take profit levels lock in profits and limit losses.

  • Large parameter optimization space Indicator parameters can be flexibly tuned to adapt to changing markets.

  • More realistic with commission factored in Commission considered makes backtest results closer to live performance.

Risk Analysis

  • Risk of misjudgments from over-optimization Too many combined indicators may also lead to misjudgments. Effectiveness needs evaluation.

  • Higher risk with short-term trading Compared to longer timeframes, 15-min is more susceptible to sudden events and risks.

  • Strategy stability needs further validation More extensive testing across long history and markets is needed to ensure reliability.

  • Lengthy optimization with multiple parameters Many parameters introduced leads to lengthy process for optimizing all parameter combinations.

Improvement Directions

  • Evaluate real effect of each indicator Backtest to verify actual incremental benefit of each indicator, avoid redundancy.

  • Optimize and test stability Test optimization results across more markets to ensure robustness.

  • Incorporate stop loss strategies Such as trailing stop, bracket order stop to further control risk.

  • Consider more cost factors Such as slippage to make backtest closer to live performance.

Summary

This strategy combines multiple indicators and risk management techniques tailored for 15-min Bitcoin trading. Large space remains for optimizing parameters, evaluating indicator effectiveness, broad market stability test, and introducing more real-world factors to find the optimal combination within the multi-factor approach. With persistent optimization and verification, it can become an effective tool for crypto high frequency trading.


/*backtest
start: 2023-08-25 00:00:00
end: 2023-09-09 00:00:00
period: 10m
basePeriod: 1m
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/

// This source code is subject to the terms of the Mozilla Public License 2.0 at https://mozilla.org/MPL/2.0/
// © deperp
//@version=5
strategy('3kilos', shorttitle='3kilos BTC 15m', overlay=true, initial_capital=100000, max_bars_back=5000, default_qty_type=strategy.percent_of_equity, default_qty_value=10, commission_type=strategy.commission.percent, commission_value=0.07, pyramiding=0)

short = input.int(50, minval=1)
srcShort = input(high, title='TEMA short')

long = input.int(100, minval=1)
srcLong = input(low, title='TEMA long 2')

long2 = input.int(350, minval=1)
srcLong2 = input(close, title='TEMA long 3')

atrLength = input.int(550, title='ATR Length', minval=1)
mult = input.float(3, title="Multiplier", minval=0.5, step=1)

smaPeriod = input.int(100, title="SMA Period", minval=1)

takeProfitPercent = input.float(1, title="Take Profit (%)", minval=0.1) / 100
stopLossPercent = input.float(3, title="Stop Loss (%)", minval=0.1) / 100


tema(src, length) =>
    ema1 = ta.ema(src, length)
    ema2 = ta.ema(ema1, length)
    ema3 = ta.ema(ema2, length)
    3 * (ema1 - ema2) + ema3

tema1 = tema(srcShort, short)
plot(tema1, color=color.new(color.red, 0), linewidth=2)

tema2 = tema(srcLong, long)
plot(tema2, color=color.new(color.blue, 0), linewidth=2)

tema3 = tema(srcLong2, long2)
plot(tema3, color=color.new(color.green, 0), linewidth=2)

// Custom ATR calculation with EMA smoothing
atr_ema(src, length) =>
    trueRange = math.max(math.max(high - low, math.abs(high - close[1])), math.abs(low - close[1]))
    emaTrueRange = ta.ema(trueRange, length)
    emaTrueRange

// Calculate ATR with EMA smoothing
atr = atr_ema(close, atrLength)

// Calculate Supertrend
var float up = na
var float dn = na
var bool uptrend = na
up := na(up[1]) ? hl2 - (mult * atr) : uptrend[1] ? math.max(hl2 - (mult * atr), up[1]) : hl2 - (mult * atr)
dn := na(dn[1]) ? hl2 + (mult * atr) : uptrend[1] ? hl2 + (mult * atr) : math.min(hl2 + (mult * atr), dn[1])
uptrend := na(uptrend[1]) ? true : close[1] > dn[1] ? true : close[1] < up[1] ? false : uptrend[1]

// Calculate SMA
sma = ta.sma(tema1, smaPeriod)

// Heikin-Ashi Close
haTicker = ticker.heikinashi(syminfo.tickerid)
haClose = request.security(haTicker, timeframe.period, close)


// Trend determination using Heikin-Ashi Close
longC = tema1 > tema2 and tema1 > tema3 and uptrend and tema1 > sma and haClose > haClose[1]
shortC = tema1 < tema2 and tema1 < tema3 and not uptrend and tema1 < sma and haClose < haClose[1]


alertlong = longC and not longC[1]
alertshort = shortC and not shortC[1]

useDateFilter = input.bool(true, title="Begin Backtest at Start Date",
     group="Backtest Time Period")
backtestStartDate = input(timestamp("1 Jan 2023"), 
     title="Start Date", group="Backtest Time Period",
     tooltip="This start date is in the time zone of the exchange " + 
     "where the chart's instrument trades. It doesn't use the time " + 
     "zone of the chart or of your computer.")

inTradeWindow = true

stopLossLevelLong = close - atr * mult
stopLossLevelShort = close + atr * mult
longTakeProfitLevel = close * (1 + takeProfitPercent)
longStopLossLevel = close * (1 - stopLossPercent)
shortTakeProfitLevel = close * (1 - takeProfitPercent)
shortStopLossLevel = close * (1 + stopLossPercent)



if inTradeWindow and longC
    strategy.entry('Long', strategy.long, comment='Long')
    strategy.exit("TP Long", "Long", limit=longTakeProfitLevel, stop=longStopLossLevel, comment="TP/SL Long")

if inTradeWindow and shortC
    strategy.entry('Short', strategy.short, comment='Short')
    strategy.exit("TP Short", "Short", limit=shortTakeProfitLevel, stop=shortStopLossLevel, comment="TP/SL Short")

// Alerts

alertcondition(longC, title='Long', message=' Buy Signal ')
alertcondition(shortC, title='Short', message=' Sell Signal ')

More