Hi Mathew,

I took a look at your algo and corrected a few issues. I attached a modified version of your algo which addresses the following issues:

1. The pipeline being attached to the algo was just an empty pipeline, it wasn't the one defined in my_pipeline (line 15). The new version attaches the output of my_pipeline to the algo.
   
2. Lines 39-44 are using the wrong namespace for fundamental data. The notation should look something like Fundamentals.market_cap.latest if you want to use the latest market cap. I fixed the calls in the attached algo.
   
3. The use of global variables in not necessary, and actually discouraged. I removed these.
   
4. There were also some deprecated functions being used, so I changed those to newer versions.

The algo still isn't ordering anything, but it's at least getting a non-empty pipeline. My suggestion as a next step would be to try debugging by printing out the state of different variables before orders are placed to make sure everything is behaving as you expect it to. I'd also recommend going through the tutorials to get a better understanding of the current API.

I hope this helps.

Hey there Jamie! Thank you very much for your help! After seeing how you fixed my algorithm, I was able to tweak the pipeline into returning a data frame. However, now I've encountered a new problem: None of my orders are going through. I will attach the back test and some code that I think is influencing this error, but I can't seem to figure out why my python code is not executing my purchases. Any help would be much appreciated.

Creates my weights for the stocks:

def create_weights(context, stocks):  
    """  
    Takes in a list of securities and calculates  
    the portfolio weighting percentage used for each stock in the portfolio  
    """  
    if len(stocks) == 0:  
        return 0  
    else:  
        # Buy only 0.95 of portfolio value to avoid borrowing  
        weight = .95/len(stocks)  
        return weight  

Next Code has to do with MACD:

def MACD(prices, fastperiod=12, slowperiod=26, signalperiod=9):  
    '''  
    Function to return the difference between the most recent  
    MACD value and MACD signal. Positive values are long  
    position entry signals 

    optional args:  
        fastperiod = 12  
        slowperiod = 26  
        signalperiod = 9

    Returns: macd - signal  
    '''  
    macd, signal, hist = talib.MACD(prices,  
                                    fastperiod=fastperiod,  
                                    slowperiod=slowperiod,  
                                    signalperiod=signalperiod)  
    return macd[-1] - signal[-1]  

Next code defines my rebalance:

def rebalance(context, data):  
    # Track cash to avoid leverage  
    cash = context.portfolio.cash  
    pricess = data.history(context.portfolio.positions, fields="price", bar_count=40, frequency="1d")  
    macd = pricess.apply(MACD, fastperiod=12, slowperiod=26, signalperiod=9)  

and finally, here is the purchase code I have:

    # Create weights for each stock  
    weight = create_weights(context, context.output)

    # Rebalance all stocks to target weight of overall portfolio  
for stock in context.output:  
        if weight != 0 and data.can_trade(stock):  
            notional = context.portfolio.portfolio_value * weight  
            price = data.current(stock, 'price')  
            numshares = int(notional / price)  
            # Growth companies could be trading thin: avoid them  
            if cash > price * numshares and numshares < data.current(stock, 'volume') * 0.2:  
                if data.can_trade(stock) and macd[stock] > 0:  
                    order_target_percent(stock, weight)  
                    cash -= notional - context.portfolio.positions[stock].amount  
                    log.info("Placing order: %s" % stock)  

Hi Matthew,

The last algo you shared is running into an issue because you are indexing macd with securities from context.output (line 129), but macd is only calculated/ only contains securities that were already in your portfolio (lines 106, 107).

To fix this, you could create a set that contains securities in your portfolio and securities from context.output, and use it to calculate macd.

I hope this helps.

Hey Ernesto, that actually did help and it made complete sense. When creating a new set for securities in context.ouput, I ran into a new error:

"Exception: inputs are all NaN There was a runtime error on line 109."

Any Idea why this is showing up??

Thank you for your help by the way, this has been a great learning experience

Hi Mathew,

Talib's functions raise that error when you pass a vector consisting entirely of NaN values as input.

The underlying issue is that HNNA has a start date of Sep, 2005, but pricing data is not available until May, 2014 (when it was listed on NASDAQ). So this asset makes it through your pipeline filters into your tradable universe before there is any pricing data for it, and data.history returns a vector of NaNs for this asset. I have made a reported of this particular issue, and I'll share it with our data team for review.

In the meantime, you can easily avoid this type of issues by adding a price filter to your pipeline screen. See the backtest attached for an example in line 46.

I hope this helps.