Trailing Stop Loss

[nigelkhakoo]

Hi - I am an experienced trader but am pretty new to coding /python. I have implemented an event driving backtest using this library with self.buy() and self.sell(). Is there anyway to add trailing stop losses to this method of backtesting or do I have to use the vectorised backtesting method which I have not yet tried ?

Great library by the way.
Thanks

[kernc]

This is what TrailingStrategy does under the hood:

backtesting.py/backtesting/lib.py

Lines 442 to 450 in a49122c

	def next(self):
	super().next()
	for trade in self.trades:
	if trade.is_long:
	trade.sl = max(trade.sl or -np.inf,
	self.data.Close[-1] - self.__atr[-1] * self.__n_atr)
	else:
	trade.sl = min(trade.sl or np.inf,
	self.data.Close[-1] + self.__atr[-1] * self.__n_atr)

Namely, it continually sets trade's .sl properties.

These strategies in backtesting.lib are meant to be independent and composable, meaning you might even be able to just inherit from them (as in this tutorial).

[nigelkhakoo]

OK got it - many thanks will try and code my own as i have a few different parameters

[kernc]

I'm curious, what other parameters do you have?

[nigelkhakoo]

Happy to discuss offline if thats ok

[atakanokan]

@kernc Since trailing SL is setting the .sl property at every bar, does this mean using SL and TSL are mutually exclusive? So it's impossible to set it up as SL @ 1% away from entry & TSL @ 2% away from entry like below?

    def next(self):
        # get previous bar close
        price = self.data["Close"][-1]

        # long/short SL and TP
        long_sl = price * 0.99
        long_tp = price * 1.01
        short_sl = price * 1.01
        short_tp = price * 0.99

        # signal logic
        if crossover(self.macd, self.macdsignal):
            self.buy(sl = long_sl, tp = long_tp)
        elif crossover(self.macdsignal, self.macd):
            self.sell(sl = short_sl, tp = short_tp)

        # tsl
        for trade in self.trades: 
            if trade.is_long: 
                trade.sl = max(trade.sl or -np.inf, 
                            self.data["Close"][-1] * 0.98) 
            else: # short
                trade.sl = min(trade.sl or np.inf, 
                            self.data["Close"][-1] * 1.02) 

[kernc]

Do you know of brokers where "trailing" is not just a modifier on the one and single SL order? Setting sl=/.sl and using TrailingStrategy is mutually exclusive, but there is nothing preventing you from coding in an arbitrary SL logic.

To have initial SL at -1% and then trail with -2%, you could:

price = self.data.Close[-1]
if buy_condition:
    self.buy(sl=price*.99)
for trade in trades:
    if price / trade.entry_price > 1.01:     # Once profit is over >1%
        trade.sl = max(trade.sl, price*.98)  # "turn on" -2% trailing SL

[atakanokan]

Not aware of any brokers but it should be possible to code the strategy on client side to achieve the same behavior. And thanks for the example script really helpful!