SuperTrend con estrategia de seguimiento de pérdidas de parada

El autor:¿ Qué pasa?, Fecha: 2023-09-18 16:05:31
Las etiquetas:

Resumen general

Esta estrategia utiliza el indicador SuperTrend para determinar la tendencia del precio e incorpora un mecanismo de stop loss para la tendencia después de las operaciones. SuperTrend se construye a partir de dos promedios móviles.

Estrategia lógica

Calcular el DEMA de líneas rápidas y lentas. La dirección de la SuperTendencia se determina a partir de la diferencia entre las dos líneas. En tendencias alcistas, el cruce de la línea rápida por encima de la línea lenta indica una entrada larga. En tendencias bajistas, el cruce de la línea rápida por debajo de la línea lenta indica una entrada corta. Después de la entrada, se establece una línea de stop loss. Si la pérdida alcanza el porcentaje de stop loss, la posición se cierra.

Ventajas

  • SuperTrend determina efectivamente la tendencia de los precios
  • Control del riesgo de pérdida por cierre de operaciones
  • Se pueden utilizar diferentes conjuntos de parámetros para diferentes períodos
  • Buena capacidad de control del descenso

Los riesgos

  • Las medias móviles se retrasan en la detección de tendencias
  • Si se establece un stop loss demasiado amplio, las pérdidas pueden aumentar.
  • La alta frecuencia del comercio aumenta los costes de transacción
  • No se pueden determinar con precisión los puntos de cambio exactos

Los riesgos pueden mitigarse acortando los períodos de media móvil, optimizando el porcentaje de pérdida de parada, reduciendo la frecuencia de las operaciones, etc.

Mejoras

  • Prueba de diferentes combinaciones de parámetros de línea rápida/lenta
  • Optimizar el porcentaje de pérdida de parada
  • Considere la posibilidad de añadir otros indicadores para determinar la inversión de tendencia
  • Prueba de robustez en diferentes productos

Conclusión

Esta estrategia simple utiliza indicadores para la dirección de la tendencia y stop loss para el control de riesgos, adecuado para el comercio intradiario.


/*backtest
start: 2023-09-10 00:00:00
end: 2023-09-13 00:00:00
period: 2m
basePeriod: 1m
exchanges: [{"eid":"Binance","currency":"BTC_USDT"}]
*/

//@version=3
strategy(title = "[Prod] SuperTrend with Stoploss+TrendZones - 1H", default_qty_type=strategy.percent_of_equity, default_qty_value=100, shorttitle = "[Prod] SuperTrend with Stoploss+TrendZones - 1H", overlay = true, commission_type=strategy.commission.percent, commission_value=0.3)
//--cyberfunk 2018 
 
////////////////////////////////////////////////////////////
// This section of code was borrowed from "Trend Trader Strategy" to 
// help us det. if we are in an uptrend or downtrend.
////////////////////////////////////////////////////////////
Length = input(12, minval=1),   // 19 / 5 also works for lenth /multiplier... but 11 /4 seems slightly more optimal somehow
Multiplier = input(4, minval=1)
avgTR      = wma(atr(1), Length)
highestC   = highest(Length)
lowestC    = lowest(Length)
hiLimit = highestC[1]-(avgTR[1] * Multiplier)
loLimit = lowestC[1]+(avgTR[1] * Multiplier)
ret = na
pos =na
ret := iff(close > hiLimit and close > loLimit, hiLimit,iff(close < loLimit and close < hiLimit, loLimit, nz(ret[1], 0)))
pos :=	iff(close > ret, 1,iff(close < ret, -1, nz(pos[1], 0))) 
//barcolor(pos == -1 ? red: pos == 1 ? green : blue )
plot(ret, color= blue , title="Trend Trader Strategy")
uptrend = if close > ret  // this will be used later to determine which domain we are in (uptrend or downtrend)
    true
else
    false
/////////////////////////////////////////////////////////////////////
// This script has been optimized for the 1 hr candle window view.  
// An older Pine2 ancestor of this script suffered from repainting issues when 
// viewing in lower resolutions than the res/2 parameter.
// running this in Pine3 mode seems to fix those problems, see https://www.tradingview.com/wiki/Pine_Script:_Release_Notes#Pine_Version_3  and https://backtest-rookies.com/2017/11/29/tradingview-indicator-repainting/ for references on the problem. 
//
//  There does seem to be ocassional repainting that affects backtrading here though, somehow.. and i'm not sure how to fix it.  V3 SHOULD have made it impossible.  Any help appreciated.
//
// Also note that the length of the DEMA sampling shouldn't be longer than a 
// candle, which could happen if you made resolution > chartres^2.  For example,
// a chartres=5min and a res=30 setting would cause sampling = 6, which is longer
// than the length of a candle (5 min, b/c chartres=5 min)
/////////////////////////////////////////////////////////////////////


fastPeriod    = input(title="fastPeriod",defval=4, minval=1)
slowPeriod = input(title="slowPeriod",defval=31, minval=1) 
resInteger = input(title="resolution",defval=98)
// The parameters above that work well highly oscillating sideways markets,
// struggle with longer-term trends, particularly strong ones, where there's the
// potential for a lot of loss or strong gain. To make sure we don't get burned 
// by long term losses, a stoploss mechanism is implemented to compensate.
// *stoplossPercent* is the point at which it becomes useful to "panic sell".
// *letitRidePercent* is the point beyond which it's more profitable NOT to panic sell.
// Like the periods, these values will have to be tuned to a particuliar market behavior.
stoplossPercent    = input(title="stopLossPercentage",type=float,defval=2, minval=1, maxval=100)
letitRidePercent = input(title="letItRidePercentage",type=float,defval=3, minval=1, maxval=100)

//To try to recapture the value of the long term strong-uptrends, first we 
// detected when we are in strong uptrend mode, by using an indicator (uptrend?)
// Then, if we are in strong uptrend mode, we use a different
// set of variables for fast/slow periods that's better suited to that situation.

uptrendFastPeriod = input(title="Uptrend FastPeriod",defval=4, minval=1) 
uptrendSlowPeriod = input(title="Uptrend SlowPeriod",defval=16, minval=1) 
//uptrendResInteger = input(title="Uptrend resolution",defval=98)

// Now that we have everything setup, for each bar we calculate the difference of emas for the downtrend/uptrend situations
res = tostring(resInteger) 
sampling =  (resInteger / timeframe.period )
demaFastNormal = security(syminfo.tickerid, res, 2 * ema(close, fastPeriod) - ema(ema(close, fastPeriod), fastPeriod))
demaSlowNormal = security(syminfo.tickerid,res, 2 * ema(close, slowPeriod) - ema(ema(close, slowPeriod), slowPeriod)  )
demaFastUT = security(syminfo.tickerid, res, 2 * ema(close, uptrendFastPeriod) - ema(ema(close, uptrendFastPeriod), uptrendFastPeriod))
demaSlowUT = security(syminfo.tickerid,res, 2 * ema(close, uptrendSlowPeriod) - ema(ema(close, uptrendSlowPeriod), uptrendSlowPeriod)  )

// now, based on our uptrend indicator, we decide which of the two line pairs to
// use to determine the crossover signals that tell us to buy/sell.
demaFast = uptrend? demaFastUT:demaFastNormal
demaSlow =  uptrend? demaSlowUT:demaSlowNormal


//  Now, we calculate the crossover signals.
// **IMPORTANT TRADING NOTE** :  You should buy/sell on the CLOSE of the
// highlighted bar with the  signal, to adhere to back-tested results!
// (Technically, the open of the NEXT bar, but these are usually effectively the same.)  
// Note that if you check "calculate on every tick", you might be decieved until 
// you reload the page, this is NOT, however, a repaint issue, as after the bar 
// after the signal closes there is a consistent buy/sell indicator at the open 
// price of the bar after the the signal, which is basically as close to same as
// the close price of the signal (red/green highlighted) bar.

buy = crossover(demaSlow[sampling], demaFast[sampling])
stoploss = ( strategy.openprofit <-1*stoplossPercent/100*strategy.equity) and not ( strategy.openprofit < -1*letitRidePercent/100*strategy.equity )
sell = strategy.opentrades!=0 and (crossunder(demaSlow[sampling], demaFast[sampling]) or crossover(ret,close) ) and not stoploss // "and not stoploss" here helps us not over-color when both conditions are true.  The stoploss is more interesting than a regular sell condition


///////////////////////////////////////
// PLOTTING SECTION 
// This is a collected place for Plotting and coloring all the things 
// for easy toggling on/off.
///////////////////////////////////////

//plot(buyPrice, title="BuyPriceLine", color = purple )   // For a study, this replaces the strategy.openprofit way to guard against stoploss
plot(demaFast)   // These are the crossover lines that trigger the buy/sell
plot(demaSlow)




bgcolor( buy ? lime : na, transp=50)
bgcolor( stoploss ? orange : na, transp=10)
bgcolor( sell ? red : na, transp=50)
bgcolor( strategy.openprofit < -1*letitRidePercent/100*strategy.equity ? purple : na, transp= 50)

bgcolor( demaSlow[sampling]> demaFast[sampling]  ? lime : na, transp=96)
bgcolor( demaSlow [sampling]< demaFast[sampling]  ? red : na, transp=99)

// the below is for backtesting from different timeframes.  By default it is set to 1/1/2017 at 1:00 AM
startdate=timestamp(2017, 01, 01, 01, 00)   // year , month, day, hour, minute

strategy.entry("BUY", strategy.long, when = buy and (time>=startdate) )
strategy.close("BUY", when = sell or stoploss)



Más.