cash not used entirely

[Ilan Seeder]

Hello,
I'm trying to do a basic project, with an all-in sizer. For some reason, there is always left-over cash which is way larger than the stock price. How can I set the trader to use all the cash up to its limits?

Here's the code:

import backtrader as bt
import IlanData

#####################################################################################
class overIndicator(bt.Indicator):
    lines = ('over',)

    def __init__(self):
        self.l.over = self.data > 170

#####################################################################################
class underIndicator(bt.Indicator):
    lines = ('under',)

    def __init__(self):
        self.l.under = self.data < 150

#####################################################################################
# margin indicator

class MarginIndicator(bt.Indicator):
    lines = ('margin',)

    def __init__(self):
        underInd = underIndicator()
        overInd = overIndicator()
        self.l.margin = overInd - underInd


#####################################################################################
# margin strategy

class MarginStrategy(bt.SignalStrategy):
    def __init__(self):
        myind = MarginIndicator()
        # bt.LinePlotterIndicator(myind.over150, name='over 150', subplot=True)
        self.signal_add(bt.SIGNAL_LONG, myind)

#####################################################################################

data = IlanData.getData()
cerebro = bt.Cerebro()
cerebro.adddata(data)
cerebro.broker.setcash(10000)
cerebro.addstrategy(MarginStrategy)
cerebro.broker.setcommission(commission = 0.0)
cerebro.addsizer(bt.sizers.AllInSizer)
cerebro.run()
cerebro.plot()

Here's the output. Note that after buying all-in, I'm left with almost 400 USD, while the cost of buying is 170.

Thanks!

[backtrader]

Since you seem to be using TSLA your code does this with a fresh download of TSLA data

The best theory: the code you posted is not the code which was executed. A hint: your volume doesn't follow the same colour variation pattern.

[backtrader]

In any case, the alternative is that your data contains huge differences from the open to the close (which could also be the explanation for the colour of the volume bars)