Must be real number, not LinesOperation

eastatlantasanta

Hi there, I'm attempting to make a custom stochastic of my own after browsing a couple on these forums. I'm gonna jump straight into the code.

class MyStochastic(bt.Indicator):
    lines = ('Stoc', 'MyStoch',)

    def __init__(self):
    
        self.f = math.degrees(1.44 * 3.14159 / 10)
        self.a1 = math.exp(-self.f)
        self.b1 = 2 * self.a1 * math.cos(self.f)
        self.c2 = self.b1
        self.c3 = -self.a1 * self.a1
        self.c1 = 1 - self.c2 - self.c3
    
        super(MyStochastic,self).__init__()
    
    def prenext(self):
        self.HighestC = bt.indicators.Highest(period=9)
        self.LowestC = bt.indicators.Lowest(period=9)
        self.l.Stoc[0] = (self.data[0]-self.LowestC/(self.HighestC-self.LowestC))
        self.lines.MyStoch[0] = self.c1*(self.l.Stoc[0]+self.l.Stoc[0])/2 + self.c2*self.l.Stoc[0] + self.c3*self.l.Stoc[0]
   
    def nextstart(self):
        self.HighestC = bt.indicators.Highest(period=9)
        self.LowestC = bt.indicators.Lowest(period=9)
        self.l.Stoc[0] = (self.data[0]-self.LowestC/(self.HighestC-self.LowestC))
        self.lines.MyStoch[0] = self.c1*(self.l.Stoc[0]+self.l.Stoc[-1])/2 + self.c2*self.l.MyStoch[-1] + self.c3*self.l.MyStoch[-1]
       
    def next(self):
        self.HighestC = bt.indicators.Highest(period=9)
        self.LowestC = bt.indicators.Lowest(period=9)
        self.l.Stoc[0] = (self.data[0]-self.LowestC/(self.HighestC-self.LowestC))
        self.lines.MyStoch[0] = self.c1*(self.l.Stoc[0]+self.l.Stoc[-1])/2 + self.c2*self.l.MyStoch[-1] + self.c3*self.l.MyStoch[-2]

Here's just a quick description of why I have it setup the way it is:
Using the three sections (prenext, nextstart, next) allow me to initialize the lines properly. This is a fix that the user @Guwop seemed unable to find. I have confirmed the effectiveness of this method with other recursive indicators before this one.

I use the super(MyStochastic,self).init() because I also utilize the Highest and Lowest indicators, and it is my understanding that under these circumstances that including super() is necessary.

Here is the error message I receive upon running...

<ipython-input-12-4e36424d155e> in prenext(self)
     75         self.HighestC = bt.indicators.Highest(period=9)
     76         self.LowestC = bt.indicators.Lowest(period=9)
---> 77         self.l.Stoc[0] = ((self.data[0]-self.LowestC)/(self.HighestC-self.LowestC))
     78         self.lines.MyStoch[0] = self.c1*(self.l.Stoc[0]+self.l.Stoc[0])/2 + self.c2*self.l.Stoc[0] + self.c3*self.l.Stoc[0]
     79 

~\Anaconda3\lib\site-packages\backtrader\linebuffer.py in __setitem__(self, ago, value)
    220             value (variable): value to be set
    221         '''
--> 222         self.array[self.idx + ago] = value
    223         for binding in self.bindings:
    224             binding[ago] = value

TypeError: must be real number, not LinesOperation

I'm not quite sure what it is that separates the values of self.l.Stoc[0] from self.l.MyStoch[0], because as I have mentioned previously, in other recursive indicators of mine, setting values to a lines object in the same manner as MyStoch has been no problem. Additionally, I have swapped out self.data[0] for an integer and receive the same error. I'd appreciate any help. Love the platform by the way. Thanks!

adrtod

Consider declaring indicators once in __init__:

self.HighestC = bt.indicators.Highest(self.data, period=9)
self.LowestC = bt.indicators.Lowest(self.data, period=9)

and then use

self.l.Stoc[0] = (self.data[0]-self.LowestC[0])/(self.HighestC[0]-self.LowestC[0])

adrtod

Alternatively in __init__ only:

HighestC = bt.indicators.Highest(self.data, period=9)
LowestC = bt.indicators.Lowest(self.data, period=9)
self.l.Stoc = (self.data-LowestC)/(HighestC-LowestC)

eastatlantasanta

@adrtod Yes that worked! Thank you very much. If it isn't too much to ask, would you mind explaining why it works that way? From what I had thought, init is a def that is only called once, and so the values for Highest/Lowest would only be calculated once, and not change with new data. Obviously this is not the case... does this mean that init is in fact called more than once?

adrtod

__init__ is called only once and only declares the indicators as objects, no computation is done during __init__ but they are handled automatically.
The same applies when using operators to combine indicators, see https://www.backtrader.com/docu/concepts.html#operators-using-natural-constructs
See also https://www.backtrader.com/docu/induse.html and https://www.backtrader.com/docu/inddev.html

eastatlantasanta

@adrtod, Thank you for the explanation it's all making sense now - especially considering how the declarative approach to indicators is setup.

backtrader

Overtaken by @adrtod. Glad to see someone with such an intimate knowledge of things.