Indicator that calculates the n-candle max of another indicator

nathanl93

Hi everyone,

I am writing an Indicator that takes another Indicator as its Data Feed. It should work generically on any Indicator that I pass it. The idea is to calculate the max out of the last n-candles and use that as the value.

Here the indicator in question:

class LastNMaxIndicator(Indicator):
    lines = ('example', )
    params = {'lookback_period': 3}

    def __init__(self):
        self.addminperiod(self.params.lookback_period)

    def next(self):
        self.lines.example[0] = max(self.data.get(size=self.params.lookback_period))

In my strategy, I am passing the SimpleMovingAverage as its input:

class TestStrategy(Strategy):
    def __init__(self):
        self.sma = MovingAverageSimple(self.datas[0])
        self.peak = LastNMaxIndicator(self.sma)

    def next(self):
        print('=== Date: {} ==='.format(self.datas[0].datetime.date(0))

However, I get an error when running this saying that the result of self.data.get(size=self.params.lookback_period)) is empty:

  File "...", line 14, in next
    self.lines.example[0] = max(self.data.get(size=self.params.lookback_period))
ValueError: max() arg is an empty sequence

What am I missing?

ab_trader

bt has Highest or MaxN indicator which returns what you want.

nathanl93

Thanks for answering. If I wanted to do it through the next method though, how would I do it?

I am not sure what I am doing wrong with my code.

nathanl93

I am asking because, in the event that there is no off-the-shelf provided functionality by Backtrader, I would like to know how to do it through the next method :)

vladisld

There is probably an issue with the minperiod calculation in your indicator. Calling the super(LastNMaxIndicator, self).__init__() may help, before adding your own minperiod.

nathanl93

@vladisld said in Indicator that calculates the n-candle max of another indicator:

There is probably an issue with the minperiod calculation in your indicator. Calling the super(LastNMaxIndicator, self).__init__() may help, before adding your own minperiod.

That was it. Thank you! For anyone wondering what the indicator looks like now:

class LastNMaxIndicator(Indicator):
    lines = ('example', )
    params = {'lookback_period': 3}

    def __init__(self):
        super(LastNMaxIndicator, self).__init__()
        self.addminperiod(self.params.lookback_period)

    def next(self):
        self.lines.example[0] = max(
            self.data.get(size=self.params.lookback_period))