self coded ATR

Gleetche

I am trying to build my own ATR, and I don't know if I coded it correctly or not, can someone confirm or point out any mistakes that I have?

https://www.investopedia.com/terms/a/atr.asp

I found a blog that teaches you how to build your own indicator, in this case which is ATR.

But the code from the blog only has current high - current close, the original atr includes:

current high - current low
current high - previous close
current low - previous close

class AverageTrueRange(bt.Strategy):

    def log(self, txt, dt=None):
        dt = dt or self.datas[0].datetime.date(0)
        print(f'{dt.isoformat()} {txt}')  # Print date and close

    def __init__(self):
        self.dataclose = self.datas[0].close
        self.datahigh = self.datas[0].high
        self.datalow = self.datas[0].low
        self.datapreviousclose = self.datas[-1].close
        self.my_atr = bt.ind.ATR(period = 14,)

    def next(self):
        range_total = 0
        for i in range(-13, 1):
            x = self.datahigh[i] - self.datalow[i]
            y = self.datahigh[i] - self.datapreviousclose[i]
            z = self.datalow[i] - self.datapreviousclose[i]
            if x > y:
                temp_truerange = x
            else:
                temp_truerange = y

            if  temp_truerange > z:
                true_range = temp_truerange
            else:
                true_range = z

            range_total += true_range
        atr = range_total / 14

        # h-l , h - previous close , l - previous close
        self.log(f'Close: {self.dataclose[0]:.2f}, ATR: {atr:.4f}')
        self.log('bt.ATR:%.2f' % (self.my_atr*1))

Gleetche

I forgot to add that, i compared the results of the builtin atr and the self-coded atr, and it is different :

run-out

@Gleetche said in self coded ATR:

self.datapreviousclose = self.datas[-1].close

I haven't looked at all your code, this jumps out. You need to use:

self.datapreviousclose = self.datas[0].close(-1)

Use round brackets to back up in init. Square brackets in next.

Gleetche

@run-out Oh, thanks!
What does self.datas[-1].close mean then?

rajanprabu

@Gleetche Isn't that what you wrote in your code ? There is no such object. In BT data indexing starts at 0 and grows progressively as one adds more data. self.datas[0].close(-1) is using the concept delayed indexing. This doc could clear lots of your doubts.

Gleetche

@rajanprabu I thought self.datas[-1].close means previous close.
thanks!

rajanprabu

datas[0] this indexing is for the data that feed to BT.. if you load two data sets then you will have datas[0] and datas[1].

datas[0].close[0] is current close for data[0] and data[0].close[-1] is previous close.