Delayed RSI Trading Strategy

Author: ChaoZhang, Date: 2023-10-07 15:38:56
Tags:

Overview

The Delayed RSI trading strategy utilizes the conventional RSI indicator to identify overbought and oversold conditions, and delays entering the market for a certain period after the signal appears to avoid unnecessary losses from fake breakouts. The main idea of this strategy is to use the RSI indicator to judge the overbought and oversold market conditions, and achieve more accurate entry timing by delaying entry based on this judgment.

Strategy Logic

This strategy uses a 21-period RSI indicator to determine overbought and oversold conditions. When the RSI indicator crosses above the user-defined overbought level (default 60), the market is considered overbought. When the RSI crosses below the user-defined oversold level (default 40), the market is considered oversold.

After identifying overbought or oversold signals, the strategy does not enter the market immediately. Instead, it starts counting the delay period. When the delay period (default 15 bars) is met, it enters short based on overbought signal, and long based on oversold signal.

The strategy allows users to adjust the delay period to achieve different entry timing. Longer delay period can avoid more fake breakouts, but may also miss better entry opportunities. Users need to adjust the delay period parameter based on the characteristics of specific products.

In addition, the strategy also implements options like stop loss, take profit, reverse trading, etc. Users can choose fixed stop loss, trailing stop loss, fixed take profit and so on to manage positions. The trading logic can also be reversed, i.e. go long on overbought signals and short on oversold signals.

Advantages

  1. Utilize RSI indicator to identify overbought/oversold conditions accurately and catch reversal opportunities. RSI is a mature oscillator widely used to identify reversals.

  2. Delayed entry avoids losses from fake breakouts. Many breakouts do not necessarily lead to real reversals. Delayed entry verifies the validity.

  3. Adjustable delay period allows accurate entry timing. Users can optimize delay period based on product characteristics for best entry.

  4. Implement stop loss and take profit to control risks. The strategy offers multiple ways like fixed SL/TP, trailing SL etc to manage risks.

  5. Reverse trading option adapts to different products. Users can choose normal or reverse logic to hedge uncertainties.

Risks

  1. Risk of fake signals from RSI. RSI signals may not always be accurate and can give false signals sometimes.

  2. Risk of missing opportunities if delay is too long. Excessive delay period can result in missed entry points.

  3. Increased loss risk from reverse trading. Although reverse trading hedges uncertainties, it may also amplify total loss.

  4. Risk of trailing SL being too close and getting stopped out prematurely.

  5. Insufficient profit due to inaccurate fixed TP. Fixed TP cannot achieve max profit and needs reasonable forecast.

To address these risks, optimization suggestions are:

  1. Filter RSI signals with other indicators like KDJ, MACD etc to improve reliability.

  2. Backtest with historical data to find optimal delay period for each product. One size does not fit all.

  3. Use reverse logic cautiously, preferably combine with trend following.

  4. Keep wide buffer for trailing SL to avoid prices getting too close.

  5. Test different TP ratios to find optimum. Consider dynamic take profit also.

Optimization Opportunities

The strategy can be further optimized in the following aspects:

  1. Combine multiple indicators to filter entry signals, e.g. KDJ, MACD with RSI for more robust signals.

  2. Dynamically adjust delay period based on market volatility. This maintains false breakout avoidance while improving entry accuracy.

  3. Optimize SL/TP strategies, such as dynamic SL, profit retracement ratio SL, time-based SL etc., making them adapt better to market swings.

  4. Incorporate trend. Gauge if breakout direction aligns with major trend. Also adjust delay period based on breakout momentum.

  5. Use machine learning to find optimal parameter combinations. ML can auto tune parameters based on large training and backtest datasets.

In conclusion, the strategy has ample room for optimization via indicator combos, dynamic tuning of parameters, trend integration etc. ML is also a promising direction going forward.

Summary

The Delayed RSI strategy overall utilizes RSI to identify overbought/oversold conditions, and delays entry for a period after signal occurs to avoid unnecessary losses from fakeouts. The strategy has advantages like accurate signal identification, delayed entry to avoid false breaks, adjustable delay period, SL/TP implementation etc. But risks like unreliable RSI signals, missed opportunities from excess delay exist. These can be improved further through optimizing signal accuracy via indicator combos, dynamic delay period tuning, better SL/TP strategies etc. The strategy has broad optimization opportunities and is worth exploring.


/*backtest
start: 2023-01-01 00:00:00
end: 2023-10-06 00:00:00
period: 1d
basePeriod: 1h
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/
// © tweakerID and © BacktestRookies

// This strategy uses a 21 period RSI with an overbought (RSI indicator 
// is greater than) level of 60 (user defined) to determines long entries and an oversold 
// (RSI indicator is less than) level of 40 (user defined) for shorts. It introduces a bar delay that starts
// counting when the RSI < Oversold or RSI > Overbought conditions are true, delaying the entry with 
// the amount of bars determined by the user. The trading logic can be reversed, which seems to work better.

//@version=4
strategy("Delayed RSI Strategy", 
     overlay=false, 
     default_qty_type=strategy.percent_of_equity, 
     default_qty_value=100, 
     initial_capital=10000, 
     commission_value=0.04, 
     calc_on_every_tick=false, 
     slippage=0)
     
direction = input(0, title = "Strategy Direction", type=input.integer, minval=-1, maxval=1)
strategy.risk.allow_entry_in(direction == 0 ? strategy.direction.all : 
 (direction < 0 ? strategy.direction.short : strategy.direction.long))

// Bought and Sold Boolean Signal
bought = strategy.position_size > strategy.position_size[1] 
 or strategy.position_size < strategy.position_size[1]

/////////////////////// STRATEGY INPUTS ////////////////////////////////////////
title1=input(true, "-----------------Strategy Inputs-------------------")  

rsiLen=input(21, title="RSI Length")
i_OB = input(60, title="Overbought")
i_OS = input(40, title="Oversold")
i_delay = input(15, title="Entry Delay (# of Bars)")
i_Close= input(false, title="Use Strategy Close")

/////////////////////// BACKTESTER /////////////////////////////////////////////
title2=input(true, "-----------------General Inputs-------------------")  

// Backtester General Inputs
i_SL=input(true, title="Use Stop Loss and Take Profit")
TS=input(false, title="Use Trailing Stop")
i_SLType=input(defval="ATR Stop", title="Type Of Stop", options=["Strategy Stop", "Swing Lo/Hi", "ATR Stop"])
i_SPL=input(defval=10, title="Swing Point Lookback")
i_PercIncrement=input(defval=3, step=.1, title="Swing Point SL Perc Increment")*0.01
i_ATR = input(14, title="ATR Length")
i_ATRMult = input(3, step=.1, title="ATR Multiple")
i_TPRRR = input(2, step=.1, title="Take Profit Risk Reward Ratio")
DPR=input(false, "Allow Direct Position Reverse")
reverse=input(true, "Reverse Trades")

// Swing Points Stop and Take Profit
SwingStopProfit() =>
    LL=(lowest(i_SPL))*(1-i_PercIncrement)
    HH=(highest(i_SPL))*(1+i_PercIncrement)
    LL_price = valuewhen(bought, LL, 0)
    HH_price = valuewhen(bought, HH, 0)
    entry_LL_price = strategy.position_size > 0 ? LL_price : na 
    entry_HH_price = strategy.position_size < 0 ? HH_price : na 
    tp=strategy.position_avg_price + (strategy.position_avg_price - entry_LL_price)*i_TPRRR
    stp=strategy.position_avg_price - (entry_HH_price - strategy.position_avg_price)*i_TPRRR
    [entry_LL_price, entry_HH_price, tp, stp]

// ATR Stop
ATRStop() =>
    ATR=atr(i_ATR)*i_ATRMult
    ATRLong = ohlc4 - ATR
    ATRShort = ohlc4 + ATR
    ATRLongStop = valuewhen(bought, ATRLong, 0)
    ATRShortStop = valuewhen(bought, ATRShort, 0)
    LongSL_ATR_price = strategy.position_size > 0 ? ATRLongStop : na 
    ShortSL_ATR_price = strategy.position_size < 0 ? ATRShortStop : na 
    ATRtp=strategy.position_avg_price + (strategy.position_avg_price - LongSL_ATR_price)*i_TPRRR
    ATRstp=strategy.position_avg_price - (ShortSL_ATR_price - strategy.position_avg_price)*i_TPRRR
    [LongSL_ATR_price, ShortSL_ATR_price, ATRtp, ATRstp]
    
// Strategy Stop
StrategyStop(bought) =>
    float LongStop = na
    float ShortStop = na
    float StratTP = na
    float StratSTP = na
    [LongStop, ShortStop, StratTP, StratSTP]

//TrailingStop
TrailingStop(SL,SSL) =>
    dif=(valuewhen(strategy.position_size>0 and strategy.position_size[1]<=0, high,0))
     -strategy.position_avg_price
    trailOffset     = strategy.position_avg_price - SL
    var tstop = float(na)
    if strategy.position_size > 0
        tstop := high- trailOffset - dif
        if tstop<tstop[1]
            tstop:=tstop[1]
    else
        tstop := na
    StrailOffset     = SSL - strategy.position_avg_price
    var Ststop = float(na)
    Sdif=strategy.position_avg_price-(valuewhen(strategy.position_size<0 
     and strategy.position_size[1]>=0, low,0))
    if strategy.position_size < 0
        Ststop := low+ StrailOffset + Sdif
        if Ststop>Ststop[1]
            Ststop:=Ststop[1]
    else
        Ststop := na
    [tstop, Ststop]
  
//Stop Loss & Take Profit Switches  
SLTPLogic(LongStop, ShortStop, StratTP, StratSTP, LongSL_ATR_price, ShortSL_ATR_price, ATRtp, ATRstp,
 entry_LL_price, entry_HH_price, tp, stp) =>
    SL= i_SLType == "Swing Lo/Hi" ? entry_LL_price : i_SLType == "ATR Stop" ? LongSL_ATR_price : LongStop
    SSL= i_SLType == "Swing Lo/Hi" ? entry_HH_price : i_SLType == "ATR Stop" ? ShortSL_ATR_price : ShortStop
    TP= i_SLType == "Swing Lo/Hi" ? tp : i_SLType == "ATR Stop" ? ATRtp : StratTP
    STP= i_SLType == "Swing Lo/Hi" ? stp : i_SLType == "ATR Stop" ? ATRstp : StratSTP
    [SL, SSL, TP, STP]


/////////////////////// STRATEGY LOGIC /////////////////////////////////////////

rsi = rsi(close, rsiLen)
isOB= rsi > i_OB
isOS= rsi < i_OS
BarsSinceOB = barssince(not isOB)
BarsSinceOS = barssince(not isOS)

BUY = BarsSinceOS == i_delay
SELL = BarsSinceOB == i_delay

/////////////////////// FUNCTION CALLS /////////////////////////////////////////

// Stops and Profits
[entry_LL_price, entry_HH_price, tp, stp] = SwingStopProfit()
[LongSL_ATR_price, ShortSL_ATR_price, ATRtp, ATRstp] = ATRStop()
[LongStop, ShortStop, StratTP, StratSTP] = StrategyStop(bought)
[SL, SSL, TP, STP] = SLTPLogic(LongStop, ShortStop, StratTP, StratSTP, 
 LongSL_ATR_price, ShortSL_ATR_price, ATRtp, ATRstp, entry_LL_price, entry_HH_price, tp, stp)
[tstop, Ststop] = TrailingStop(SL,SSL)

// Entries
if reverse
    if not DPR
        strategy.entry("long", strategy.long, when=SELL and strategy.position_size == 0)
        strategy.entry("short", strategy.short, when=BUY and strategy.position_size == 0)
    else     
        strategy.entry("long", strategy.long, when=SELL)
        strategy.entry("short", strategy.short, when=BUY)
else
    if not DPR 
        strategy.entry("long", strategy.long, when=BUY and strategy.position_size == 0)
        strategy.entry("short", strategy.short, when=SELL and strategy.position_size == 0)
    else
        strategy.entry("long", strategy.long, when=BUY)
        strategy.entry("short", strategy.short, when=SELL)
// Exits
if i_SL
    strategy.exit("TP & SL", "long", limit=TP, stop=TS? tstop : SL)
    strategy.exit("TP & SL", "short", limit=STP, stop=TS? Ststop : SSL)
    
if i_Close
    strategy.close_all(when=cross(rsi, 50))

/////////////////////// PLOTS //////////////////////////////////////////////////

//Plots
rsiplot = plot(rsi, "RSI", color=#7E57C2)
band1 = hline(i_OB, "Upper Band", color=#787B86)
bandm = hline(50, "Middle Band", color=color.new(#787B86, 50))
band0 = hline(i_OS, "Lower Band", color=#787B86)
fill(band1, band0, color=color.rgb(126, 87, 194, 90), title="Background")
plot(rsi, "RSI", color=#7E57C2)
// OSOBCount = plot(isOB ? BarsSinceOB : isOS ? BarsSinceOS : na, transp=100)
// OSOBColor = color.from_gradient(isOB ? BarsSinceOB : BarsSinceOS, 0, 20, color.black, isOB ? color.red : isOS ? color.green : na)
// OBP = plot(rsi > i_OB ? rsi : na, color=color.white, display=display.none)
// fill(plot(i_OB, display=display.none), OBP, color=OSOBColor, transp=0, fillgaps=false)
// OSP = plot(rsi < i_OS ? rsi : na, color=color.white, display=display.none)
// fill(plot(i_OS, display=display.none), OSP, color=OSOBColor, transp=0, fillgaps=false)

// plotshape(BUY ? 1 : na, style=shape.arrowdown, location=location.bottom, 
//  color=color.green, title="Bullish Setup", size=size.normal)
// plotshape(SELL ? 1 : na, style=shape.arrowup, location=location.top, 
//  color=color.red, title="Bearish Setup", size=size.normal)



More