Oscilação de impulso cruzando bandas de Bollinger com estratégia de média móvel

Autora:ChaoZhang, Data: 2023-12-19 11:34:46
Tags:

Resumo

Trata-se de uma estratégia quantitativa de negociação baseada em Bollinger Bands e indicadores MACD, que combina dois principais indicadores técnicos para identificar oportunidades de negociação, com o objetivo de alcançar uma maior taxa de ganho em mercados de tendência.

A estratégia estabelecerá uma posição longa quando o preço atravessa a faixa inferior das Bandas de Bollinger para seguir a tendência e uma posição próxima quando o preço atravessa a faixa superior.

Estratégia lógica

A estratégia consiste principalmente em bandas de Bollinger e indicadores MACD.

As bandas de Bollinger calculam as bandas superior e inferior com base no desvio padrão dos preços. A quebra ascendente da banda superior sinaliza a condição de sobrecompra, enquanto a quebra descendente da banda inferior sinaliza a condição de sobrevenda. Esta estratégia é longa quando o preço quebra a banda inferior e fecha a posição quando quebra a banda superior.

O indicador MACD julga o impulso e a direção dos preços. O cruzamento da média móvel de curto prazo acima da média móvel de longo prazo é um sinal de compra, enquanto o cruzamento abaixo é um sinal de venda.

Além disso, o indicador RSI pode ajudar a identificar os níveis de sobrecompra / sobrevenda.

Vantagens da estratégia

A estratégia combina as Bandas de Bollinger, os indicadores MACD e RSI, que podem determinar efetivamente a tendência e a volatilidade dos preços.

Bandas de Bollinger capturam a tendência que se segue quando o preço sai das bandas
O MACD filtra falsos sinais das Bandas de Bollinger julgando o momento
O RSI evita comprar no pico identificando níveis de sobrecompra/supervenda
A taxa de vitória mais elevada pode ser alcançada através da otimização de parâmetros

Riscos da Estratégia

Há também alguns riscos a tomar em consideração:

Risco elevado de stop loss quando os preços flutuam violentamente
Redução da rentabilidade com ajustes de parâmetros inadequados
O MACD pode errar quando a tendência se inverte

Contramedidas:

Percentagem de stop loss pode ser afrouxada adequadamente
Requer-se um extenso backtesting para encontrar parâmetros ótimos
Mais indicadores podem ser utilizados para prever a inversão da tendência

Orientações para a otimização

As principais direcções para otimizar a estratégia incluem:

Otimizar os parâmetros das bandas de Bollinger para mais regimes de mercado
Aumentar os indicadores para melhorar a robustez
Utilize aprendizado de máquina para otimizar automaticamente parâmetros
Desempenho da estratégia de ensaio em dados de alta frequência
Adicionar módulo de gestão de risco ao limite de perdas por transação

Conclusão

No geral, esta é uma tendência típica após a estratégia. Ao combinar vários indicadores técnicos, melhora a robustez e pode alcançar uma taxa de vitória decente quando os sinais são precisos. No entanto, os riscos precisam ser monitorados. Mais melhorias podem ser feitas através de otimização e ajuste contínuos.

/*backtest
start: 2022-12-12 00:00:00
end: 2023-12-18 00:00:00
period: 1d
basePeriod: 1h
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/

//@version=4
// This source code is subject to the terms of the Mozilla Public License 2.0 at https://mozilla.org/MPL/2.0/
// © tedwardd

// This strategy is intended to help users of the 3commas.io platform backtest bot performance based on a Bollinger Strategy.
// It can also be used to signal a bot to open a deal by providing the Bot ID, email token and trading pair in the strategy settings screen.
// As currently written, this strategy uses a basic Bollinger Band strategy, recommening a deal start when the closing price crosses under the lower band.
// The thick red line plotted on the chart shows the average entry price of the current deal.

strategy("[v1.3laoowai]BNB_USDT_3m_3Commas_Bollinger_Strategy_by_tedwardd", overlay=true, default_qty_type=strategy.cash, default_qty_value=1000, initial_capital=900, currency="USD", commission_value=0.1)

// 3Commas Bot settinsg
bot_type                = input(title="Simple bot", defval="simple", options=["simple", "composite"])
bot_id                  = input(title="3Commas Bot ID", defval="")
email_token             = input(title="Bot Email Token", defval="")
base_order_size         = input(title="Base order size",minval=10, step=1, defval=10)
safety_order_size       = input(title="Safety order size", minval=15, step=1, defval=400)
volume_scale            = input(title="Safety Order Vol Scale (%)", minval=0.00, step=0.01, defval=1.83)
safety_step             = input(title="Safety Order Step Scale (%)", minval=0.00, step=0.1, defval=1.55)
safety_max              = input(title="Max Number of Safety Orders", minval=0, step=1, defval=2)
initial_deviation_input = input(title="Initial SO Deviation (%)", minval=0, step=0.01, defval=1.54) * 0.01
stoploss_input          = input(title="Long Stop Loss (%)", minval=0, step=1, defval=15) * 0.01
takeprofit_input        = input(title="Long Take Profit (%)", minval=0, step=1, defval=1.4) * 0.01

// USER INPUTS
sma_short_val           = input(title="Short MA Window", defval=21)
sma_long_val            = input(title="Long MA Window", defval=100)
ubOffset                = input(title="Upper Band Offset", defval=2.2, step=0.5)
lbOffset                = input(title="Lower Band Offset", defval=2.40, step=0.5)
cross                   = input(title="Entrry at Cross Over/Under Lower", defval="under", options=["over", "under"])

// Backtesting Date Ranges
startDate  = input(title="Start Date", defval=1, minval=1, maxval=31)
startMonth = input(title="Start Month", defval=1, minval=1, maxval=12)
startYear  = input(title="Start Year", defval=2016, minval=1800, maxval=2100)
endDate    = input(title="End Date", defval=31, minval=1, maxval=31)
endMonth   = input(title="End Month", defval=12, minval=1, maxval=12)
endYear    = input(title="End Year", defval=2022, minval=1800, maxval=2100)

// VARS
short_sma        = sma(close, sma_short_val)
long_sma         = sma(close, sma_long_val)
stdDev           = stdev(close, sma_short_val)
upperBand        = short_sma + (stdDev * ubOffset)
lowerBand        = short_sma - (stdDev * lbOffset)
stoploss_value   = strategy.position_avg_price * (1 - stoploss_input)
takeprofit_value = strategy.position_avg_price * (1 + takeprofit_input)
initial_dev_val  = strategy.position_avg_price * (1 - initial_deviation_input)
inDateRange      = true

initial_deviation = close < initial_dev_val

// Market Conditions
goodBuy    = cross=="over"?crossover(close, lowerBand):crossunder(close, lowerBand) // Buy when close crossing lower band
safety     = initial_deviation and (1-(close/strategy.position_avg_price))/.01 > strategy.opentrades-1 * safety_step and strategy.opentrades <= safety_max // SO when price deviates below SO threshold %
stoploss   = close <= stoploss_value // Stoploss condition - true if closing price for current bar drops below stoploss %
takeprofit = close >= takeprofit_value // Take profit condition - true if closing price for current bar is >= take profit percentage
goodSell = crossover(high, upperBand)

// goodSell is currently unused for any practical purpose. If you wish to try it, switch these two values. 
// Doing so will make sell suggestions at high crossover upper bollinger but it does not trigger the bot to sell as written but may affect backtest results

// Plot some lines
plot(short_sma, color=color.green)
plot(upperBand)
plot(lowerBand, color=color.yellow)
plot(strategy.position_avg_price, color=color.red, linewidth=3)


// Webhook message. Defaults to string. To signal 3c bot, fill in bot_id and email_token in user settings
var enter_msg = "Enter Position"
var exit_msg  = "Exit Position"
var close_all = "Exit Position"
if bot_id != "" and email_token != ""
    if bot_type == "composite"
        enter_msg := '{"message_type": "bot", "bot_id": ' + bot_id + ', "email_token": "' + email_token + '", "delay_seconds": 0, "pair": "' + syminfo.currency + "_" + syminfo.basecurrency + '"}'
    else
        enter_msg := '{"message_type": "bot", "bot_id": ' + bot_id + ',  "email_token": "' + email_token + '", "delay_seconds": 0}'
    if bot_type == "composite"
        exit_msg := '{"message_type": "bot", "bot_id": ' + bot_id + ', "email_token": "' + email_token + '", "delay_seconds": 0, "pair": "' + syminfo.currency + "_" + syminfo.basecurrency + '", "action": "close_at_market_price"}'
    else
        exit_msg := '{"message_type": "bot", "bot_id": ' + bot_id + ', "email_token": "' + email_token + '", "delay_seconds": 0, "action": "close_at_market_price"}'
    close_all := '{"message_type": "bot", "bot_id": ' + bot_id + ', "email_token": "' + email_token + '", "delay_seconds": 0, "action": "close_at_market_price_all"}'

actual_safety_size = float(safety_order_size) // Set safety order size to starting safety
if strategy.opentrades > 1 // If we have more than two open trades we need to start scaling the safety size by the volume_scale
    actual_safety_size := (strategy.position_size - base_order_size) * volume_scale // Remove base order from total position size and scale it for next safety order

// Momentum Strategy (BTC/USDT; 1h) - MACD (with source code) by Drun30

//@version=4
// Getting inputs
fast_length = input(title="Fast Length", type=input.integer, defval=23,group="MACD")
slow_length = input(title="Slow Length", type=input.integer, defval=16,group="MACD")
src = input(title="Source", type=input.source, defval=open,group="MACD")

signal_length = input(title="Signal Smoothing", type=input.integer, minval = 1, maxval = 50, defval = 9,group="MACD")
sma_source1 = input(title="Simple MA FAST (Oscillator)", defval="EMA", options=["HMA","DHMA","THMA","FHMA","WMA","DWMA","TWMA","FWMA","SMA","DSMA","TSMA","FSMA","EMA","DEMA","TEMA","FEMA","RMA","DRMA","TRMA","FRMA"],group="MACD")
sma_source2 = input(title="Simple MA SLOW (Oscillator)", defval="EMA", options=["HMA","DHMA","THMA","FHMA","WMA","DWMA","TWMA","FWMA","SMA","DSMA","TSMA","FSMA","EMA","DEMA","TEMA","FEMA","RMA","DRMA","TRMA","FRMA"],group="MACD")

sma_signal = input(title="Simple MA(Signal Line)",defval="EMA", options=["HMA","DHMA","THMA","FHMA","WMA","DWMA","TWMA","FWMA","SMA","DSMA","TSMA","FSMA","EMA","DEMA","TEMA","FEMA","RMA","DRMA","TRMA","FRMA"],group="MACD")
// Calculating
ma(source,length,type)=>
    type=="FEMA"?4*ema(source,length)-ema(ema(ema(ema(source,length),length),length),length):type=="FSMA"?4*sma(source,length)-sma(sma(sma(sma(source,length),length),length),length):type=="FWMA"?4*wma(source,length)-wma(wma(wma(wma(source,length),length),length),length):type=="FRMA"?4*rma(source,length)-rma(rma(rma(rma(source,length),length),length),length):type=="TEMA"?3*ema(source,length)-ema(ema(ema(source,length),length),length):type=="TSMA"?3*sma(source,length)-sma(sma(sma(source,length),length),length):type=="TWMA"?3*wma(source,length)-wma(wma(wma(source,length),length),length):type=="TRMA"?3*rma(source,length)-rma(rma(rma(source,length),length),length):type=="EMA"?ema(source,length):type=="SMA"?sma(source,length):type=="WMA"?wma(source,length):type=="RMA"?rma(source,length):type=="DEMA"?2*ema(source,length)-ema(ema(source,length),length):type=="DSMA"?2*sma(source,length)-sma(sma(source,length),length):type=="DWMA"?2*wma(source,length)-wma(wma(source,length),length):type=="DRMA"?2*rma(source,length)-rma(rma(source,length),length):type=="HMA"?hma(source,length):type=="DHMA"?2*hma(source,length)-hma(hma(source,length),length):type=="THMA"?3*hma(source,length)-hma(hma(hma(source,length),length),length):type=="FHMA"?4*hma(source,length)-hma(hma(hma(hma(source,length),length),length),length):ema(source,length)
fast_ma = ma(src,fast_length,sma_source1)  
slow_ma = ma(src,slow_length,sma_source2)
macd = fast_ma - slow_ma //Differenza tra la media mobile veloce e quella lenta 
signal = ma(macd,signal_length,sma_signal) //usa o la SMA oppure la EMA sulla differenza tra la media mobile veloce e lenta
hist = macd - signal //Differenza tra la differenza precedente e la media mobile della differenza

use_stress=input(true,title="Use stress on recent bars",group="Stress")
recent_stress=input(0.41,title="Stress on recent bars",group="Stress",step=0.01,minval=0.01,maxval=0.99)
level=input(6,title="Level of stress",group="Stress")
if use_stress 
    macd:=macd*(1/(1-recent_stress))
    if not na(macd[1])
        macd:=pow((macd*(recent_stress)),level)+(1-recent_stress*macd[1])

use_ma= input(true,title="Use moving average (MACD)?",group="Moving Average")
if use_ma
    macd:=ma(macd,input(36,title="Length",group="Moving Average"),input(title="Type MA",defval="THMA", options=["HMA","DHMA","THMA","FHMA","WMA","DWMA","TWMA","FWMA","SMA","DSMA","TSMA","FSMA","EMA","DEMA","TEMA","FEMA","RMA","DRMA","TRMA","FRMA"],group="Moving Average"))

use_linreg= input(true,title="Use linear regression (MACD)?",group="Linear Regression")
if use_linreg
    macd:=linreg(macd,input(10,title="Length",group="Linear Regression"),input(1,title="Offset",group="Linear Regression"))

//macd == linea blu (differenza tra media mobile veloce e media mobile lenta)
//signal == linea arancione (media mobile dell'macd)
//hist == istogramma (differenza tra macd e media mobile)

on_cross = input(false,title="Use cross macd and signal",group="Condition entry/exit")
on_minmax = input(true,title="Use min/max macd",group="Condition entry/exit")


aperturaLong = change(macd)>0//crossover(macd,signal)
aperturashort=not (change(macd)>0)//crossunder(macd,signal)

if on_cross
    on_minmax:=false
    aperturaLong := crossover(macd,signal)
    aperturashort := crossunder(macd,signal)
if on_minmax
    on_cross:=false
    aperturaLong := change(macd)>0//crossover(macd,signal)
    aperturashort:=change(macd)<0//crossunder(macd,signal)

rsiFilter = input(false,title="Use RSI filter?",group="RSI")
rsiTP = input(true,title="Use RSI Take Profit?",group="RSI")

len=input(22,title="RSI period",group="RSI")
srcr=input(close,title="RSI source",group="RSI")
rsi=rsi(srcr,len)
ovb=input(90,title="Overbought height",group="RSI") 
ovs=input(45,title="Oversold height",group="RSI")
okLong=rsi<ovb and change(macd)>0 and change(macd)[1]<=0
okShort=rsi>ovs and change(macd)<0 and change(macd)[1]>=0
if not rsiFilter
    okLong:=true
    okShort:=true
    
usiLong=input(true,title="Use long?")
usiShort=input(true,title="Use short?")

chiusuraShort=rsi<ovs or (aperturaLong)
chiusuraLong=rsi>ovb or (aperturashort)
if rsiTP
    aperturaLong := change(macd)>0 and change(macd)[1]<=0 and rsi<ovb//crossover(macd,signal)
    aperturashort:=change(macd)<0 and change(macd)[1]>=0 and rsi>ovs//crossunder(macd,signal)

if not rsiTP
    chiusuraShort:=okLong and aperturaLong
    chiusuraLong:=okShort and aperturashort
    
//if chiusuraShort 
//    strategy.close("SHORTISSIMO")
//if usiLong and strategy.position_size<=0 and okLong and aperturaLong
//    strategy.entry("LONGHISSIMO",true)
//if chiusuraLong 
//    strategy.close("LONGHISSIMO")
//if usiShort and strategy.position_size>=0 and okShort and aperturashort
//    strategy.entry("SHORTISSIMO",false)

// Strategy Actions
//Buy
if inDateRange and goodBuy
    strategy.entry("Good Buy", strategy.long, base_order_size, when = strategy.opentrades <= 0, alert_message=enter_msg)
if inDateRange and safety
    strategy.order("Good Buy", strategy.long, actual_safety_size, when = strategy.opentrades > 0, comment = "safety order", alert_message=enter_msg)

// Sell
if inDateRange and goodSell
    strategy.close_all(comment="Good sell point", alert_message=exit_msg)
if inDateRange and stoploss
    strategy.close_all(comment="Stoploss", alert_message=exit_msg)
//if inDateRange and takeprofit
//    strategy.close_all(comment="TP Target", alert_message=exit_msg)
if usiShort and strategy.position_size>=0 and okShort and aperturashort
    strategy.close_all(comment="SHORTISSIMO", alert_message=exit_msg)
//if chiusuraShort
//    strategy.close_all(comment="SHORTISSIMO1")

Mais.