Strategy vault / Trend following
RSI Trend Following Strategy with Trailing Stop Loss 440544
This is a quantitative trading strategy that utilizes the RSI indicator to determine the market trend and sets stop loss and take profit to lock in profits and minimize risks.

Overview
This is a quantitative trading strategy that utilizes the RSI indicator to determine the market trend and sets stop loss and take profit to lock in profits and minimize risks.
Strategy Logic
The strategy mainly uses the RSI indicator to determine the market trend direction for long or short trades. When the RSI line crosses above the lower line, it is determined as an upward trend and goes long. When the RSI line crosses below the upper line, it is judged as a downward trend and goes short.
At the same time, the strategy tracks the entry price of each order and sets a floating stop loss and take profit. For long orders, a certain percentage of the entry price is set as the stop loss line, and for short orders, a certain percentage of the entry price is set as the take profit line. When the price hits the stop loss or take profit line, the position will be closed automatically.
Advantages
- Use RSI indicator to determine market trend, avoid trading in range-bound markets;
- Set floating stop loss and take profit to lock in profits flexibly and control risks effectively;
- RSI parameters and stop loss/take profit ratios are adjustable for optimization.
Risks
- RSI indicator has some lagging, may miss short-term trend reversal points;
- Stop loss and take profit lines that are too close may get hit easily.
Optimization
- Test RSI indicators with different periods;
- Test different parameter combinations to find the optimal stop loss/take profit ratios;
- Add additional indicators for signal filtering.
Conclusion
In summary, this is a quantitative trading strategy that uses RSI indicator to track trends and incorporate floating stop loss and take profit. Compared to single indicator strategies, this strategy manages risks quite well by locking in profits flexibly. Further improvements can be made through parameters optimization and adding auxiliary indicators.
Parameters
| Name | Default | What it controls |
|---|---|---|
| v_input_int_1 | 8 | (?Signals)5 day input or RSI1 |
| v_input_int_2 | 28 | RSI Lower |
| v_input_int_3 | 72 | RSI Upper |
| v_input_int_4 | 200 | EMA Period |
| v_input_int_5 | true | Look back days for close/open |
| v_input_int_6 | 5 | (?Order Controls)max open orders |
| v_input_int_7 | 40 | Buy breakout range |
| v_input_float_1 | 1.15 | (?TPSL)Buy TP: 1+TP %, .05 seems to work well. |
| v_input_float_2 | 0.975 | Sell TP: 1-TP%. .025 seems to work well. |
| v_input_bool_1 | false | (?Alerts)Turns on Buy/Sell Alerts |
| v_input_bool_2 | false | Use default Buy/Sell or the messages below |
| v_input_1 | Buy | Buy signal API/TXT message template |
| v_input_2 | Sell | Sell signal API/TXT message template |
Source Pine Script
/*backtest
start: 2023-12-01 00:00:00
end: 2023-12-31 23:59:59
period: 1h
basePeriod: 15m
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/
// ©chewyScripts.
//@version=5
strategy("96er RSI+200EMA Strategy + Alerts", overlay=true, shorttitle = "The old 96er - RSI5 + 200 EMA")
//,use_bar_magnifier=false
// This works best on a small account $100, with 50% of equity and up to 10 max open trades.
// 96% Profitable, turns $100 into $350 in 1 month. very few losses. super happy with it.
// So far it triples the account on a 1m chart in 1 month back testing on the SEI-USD pair.
// I did not test on FX pairs or other instruments.
// had some issues with the inputs not working so had to hard code some, also the lastClose var sometimes breaks and starts following every candle, not sure why.
in_r1 = input.int(8,"5 day input or RSI1", group = "Signals")
in_lowerRSI = input.int(28,"RSI Lower", group = "Signals")
in_upperRSI = input.int(72,"RSI Upper ", group = "Signals")
in_emaperiod = input.int(200,"EMA Period", group = "Signals")
in_daysback = input.int(1,"Look back days for close/open", group = "Signals")
in_openOrders = input.int(5,"max open orders",tooltip = "Be careful, to high and you will get margin called!! 5 is probably the highest you should go", group = "Order Controls")
in_buybreakout = input.int(40,"Buy breakout range", group = "Order Controls")
in_buyTP = input.float(1.1500,"Buy TP: 1+TP %, .05 seems to work well.", group = "TPSL")
in_sellTP = input.float(0.9750, "Sell TP: 1-TP%. .025 seems to work well. ", group = "TPSL")
in_useAlerts = input.bool(false,"Turns on Buy/Sell Alerts",group = "Alerts")
in_useCustomAlertMSG = input.bool(false,"Use default Buy/Sell or the messages below",group = "Alerts")
in_alertBuySignalTxt = input("Buy","Buy signal API/TXT message template", tooltip = "Review the UserGuid on JSON varibles in alerts", group = "Alerts")
in_alertSellSignalTxt = input("Sell","Sell signal API/TXT message template", tooltip = "Review the UserGuid on JSON varibles in alerts", group = "Alerts")
simple int rsi5 = in_r1
// 3 rsi strategy , when all of them are overbought we sell, and vice versa
rsi7 = ta.rsi(close,rsi5)
[lastOpen, lastClose] = request.security(syminfo.tickerid, "D", [open,close], lookahead = barmerge.lookahead_on)
rsi3 = ta.rsi(close[5],rsi5)
ma = ta.ema(close,in_emaperiod)
plot(rsi7,"5 Day RSI",color.red)
plot(lastClose,"Previous Days Close",color.green)
plot(lastOpen,"Previous Days Open",color.white)
plot(rsi3,"Previous 5th candles RSI",color.purple)
plot(ma,"200 EMA",color.blue)
//sell = ta.crossunder(rsi7,70) and ta.crossunder(rsi14,70) and ta.crossunder(rsi21,70)
//buy = ta.crossover(rsi7,in_lowerRSI) and close < ma and rsi3 <= in_upperRSI and strategy.opentrades < in_openOrders
//sell = ta.crossunder(rsi7,in_upperRSI) and close > ma and rsi3 >= in_lowerRSI3 and strategy.opentrades < in_openOrders
//buy condition
buy = ta.crossover(rsi7,in_lowerRSI) and close < ma and close < lastClose and strategy.opentrades < in_openOrders
// sell condition
sell = ta.crossunder(rsi7,in_upperRSI) and close > ma and close > lastClose and strategy.opentrades < in_openOrders
var lastBuy = close
var lastSell = close
//var buyLabel = label.new(na,na,yloc = yloc.belowbar, style = label.style_none, textcolor = color.green, size = size.normal)
//var sellLabel = label.new(na,na,yloc = yloc.abovebar, style = label.style_none, textcolor = color.red, size = size.normal)
if (buy)
strategy.entry("BUY", strategy.long,alert_message = "Buy @"+str.tostring(close))
lastBuy := close
//buyLabel := label.new(na,na,yloc = yloc.belowbar, style = label.style_none, textcolor = color.green, size = size.normal)
//label.set_x(buyLabel,bar_index)
//label.set_y(buyLabel,low)
//label.set_text(buyLabel,"Buy!!@ " +str.tostring(lastBuy) + "\n TP: " + str.tostring(lastBuy*in_buyTP) + "\n↑")
if(not in_useAlerts)
alert("Buy")
//label.delete(buyLabel)
if ((close >= lastBuy*in_buyTP ) or (rsi7 > in_buybreakout) and close >= lastClose and (close >= lastClose*in_buyTP or close >= lastBuy*in_buyTP ) )
//label.new(bar_index,na,"TP!!@ " +str.tostring(close), yloc = yloc.abovebar, style = label.style_none, textcolor = color.green, size = size.normal)
strategy.close("BUY", "BUY Exit",alert_message = "Buy Exit: TP @" +str.tostring(close) + " OR TP: " + str.tostring(lastBuy*in_buyTP))
if(not in_useAlerts)
alert("Buy Exit")
if (sell)
strategy.entry("SELL", strategy.short, alert_message = "Sell @ " + str.tostring(close))
lastSell := close
//sellLabel := label.new(na,na,yloc = yloc.abovebar, style = label.style_none, textcolor = color.red, size = size.normal)
//label.set_x(sellLabel,bar_index)
//label.set_y(sellLabel,high)
//label.set_text(sellLabel,"Sell!!@ " +str.tostring(lastSell) + "\n TP: " + str.tostring(lastSell*in_sellTP) + "\n?")
if(not in_useAlerts)
alert("Sell")
//label.delete(sellLabel)
if ( close < ma and (close <= lastSell*in_sellTP ) or (close < lastClose*in_sellTP) )
//label.new(bar_index,na,"TP!!@ " +str.tostring(close), yloc = yloc.belowbar, style = label.style_none, textcolor = color.red, size = size.normal)
strategy.close("SELL", "Sell Exit", alert_message = "Sell Exit TP @" +str.tostring(close) + " OR TP: " + str.tostring(lastSell*in_sellTP))
if(not in_useAlerts)
alert("Sell Exit")
alertcondition(buy and in_useAlerts,"Buy Alert","test")Published by ChaoZhang on the FMZ strategy square · last modified 2024-01-31. Collected in the Quant Trading Vault. Unvalidated research: re-test with your own data and costs before trusting it.
Build this with AgenKit
A spec is not a trading system. Paste this into your coding agent and AgenKit runs a specialist team — architecture, test-first build, risk gates, review — with paper trading on by default.
/agenkit build a production-grade trading bot with a backtest engine, risk controls and paper trading by default from strategies/RSI-Trend-Following-Strategy-with-Trailing-Stop-Loss-440544.md- npx agenkit install engineering-kit in your project
- Paste the command into your agent and approve each phase
- Get tests, backtest, risk controls and a paper-trading switch
Understand it first
Before you run this, read the lesson behind the idea and how to backtest it honestly. Specs like this one are the survivors of an unknown number of trials; deflate their results accordingly.
Related specs
- Adaptive ATR and RSI Trend Following Strategy with Trailing Stop LossPine Script · Trend following · 1h
- Advanced Multi Indicator Multi Dimensional Trend Cross Quantitative StrategyPine Script · Trend following · 3h
- Bollinger Bands and RSI Trend Following StrategyPine Script · Trend following · 1h
- Dual ATR Channel Trend Following StrategyPine Script · Trend following · 15m
- Dual EMA Crossover with RSI Trend Confirmation StrategyPine Script · Trend following · 3h
- Dual Mode Adaptive Trend Trading Strategy EMA Crossover with ATR Volatility Based Risk Management SystemPine Script · Trend following · 1d