I think that's a pretty nifty little demo, thank you.

I threw in a "cash" line to show why the two prices aren't the same - the Quantopian version is getting dividends as a cash payment, not a price adjustment like Yahoo Finance does it.

I also changed the date range to run to yesterday. If you scroll all the way to the bottom of the logs, you'll see that the prices finally match up for yesterday, which you'd expect - there is no price adjustment on yesterday's price.

On a related note, I'm having a lot of trouble importing dividend history into the IDE. I'm getting the error: AttributeError: 'SIDData' object has no attribute 'datetime'. What am I doing incorrectly here?

def rename_col(df):  
    df = df.rename(columns={'Dividends': 'div'})  
    df = df.fillna(method='ffill')  
    #df = df[['price', 'sid']]  
    return df

@batch_transform(window_length=20)  
def get_past_prices(data):  
    prices = data['div']  
 #   Yprices = data['spy']  
    return prices

def initialize(context):  
    # this will add a column 'ibm' to data containing Yahoo dividend yields  
    fetch_csv('http://ichart.finance.yahoo.com/table.csv?s=IBM&a=00&b=2&c=1962&d=06&e=9&f=2013&g=v&ignore=.csv',  
        date_column='Date',  
        date_format='%Y-%m-%d',  
        symbol='ibm',  
        usecols=['Dividends'],  
        post_func=rename_col)  
    context.IBM = sid(3766);  
    context.day = 0

def handle_data(context, data):  
    ibm_data = data['ibm']  
    if context.day == 0 :  
        order (context.IBM, 10)  
        context.day = 1  
    # batch prices  
    all_prices = get_past_prices(data)  
    if all_prices is None :  
        return  
    else :  
        log.info('\n all_prices\n  %s' % all_prices)    

    # plot it  
    record(YSPY=ibm_data['div'], QSPY=data[context.IBM]['price'])  
    log.info('\nY IBM = %s' % ibm_data['div'] + '  Q IBM = %s' % data[context.IBM]['price'])  

Jim, next time feel free to create a new thread ;)

Your file only has two columns in it, Date and Dividends. You need to have another column, one that includes the symbol (docs here). I see that you're trying to indicate that the symbol is IBM, but you need to get IBM loaded into the datapanel. I bet there is a way to do that in your pre_func, but I'm not enough of a pandas expert to say how to do that.

Also, you understand that the Quantopian backtester already includes dividends in the backtest returns?