Issues in run a strategy

Pedro Ivo Paulo da Conceição

Hey guys, i'm new here and learning how to use this lib. I'm trying use somme of Teddy Koker's strategies to test a particular data and i'm having problems in some part of the code.

here is the original code: https://teddykoker.com/2019/05/improving-cross-sectional-mean-reversion-strategy-in-python/

the problem is in this part of code when try put the indicators:

def max_n(array, n):
    return np.argpartition(array, -n)[-n:]


class CrossSectionalMR2(bt.Strategy):
    params = (
        ('num_positions', 100),
    )

    def __init__(self):
        self.inds = {}
        for d in self.datas:
            self.inds[d] = {}

#the problem is in the following line

            self.inds[d]["pct"] = bt.indicators.PercentChange(d.close, period=1)

    def prenext(self):
        self.next()

    def next(self):
        available = list(filter(lambda d: len(d), self.datas))  # only look at data that existed yesterday
        rets = np.zeros(len(available))
        for i, d in enumerate(available):
            rets[i] = self.inds[d]['pct'][0]

        market_ret = np.mean(rets)
        weights = -(rets - market_ret)
        max_weights_index = max_n(np.abs(weights), self.params.num_positions)
        max_weights = weights[max_weights_index]
        weights = weights / np.sum(np.abs(max_weights))

        for i, d in enumerate(available):
            if i in max_weights_index:
                self.order_target_percent(d, target=weights[i])
            else:
                self.order_target_percent(d, 0)

I simply copied the and run in the pycharm. So i got this:

ab_trader

The issue is in the max_n function based on the error report. This is not related to bt, so you may want to debug it and check what is going on.

ab_trader

@Pedro-Ivo-Paulo-da-Conceição said in Issues in run a strategy:

i'm new here and learning how to use this lib.

I would recommend you to start with the Quickstart Guide and samples in the Docs in this case, not with the fancy complex solutions of advanced coders.