You want the return in the 120 days?

You could do something like?

    price_history = history(bar_count=days, frequency="1d", field='price')  
    p_price = p_price / p_price[0] - 1  
    p_price = np.sum(p_price)  

There is also a magic function

 price_history = history(bar_count=days, frequency="1d", field='price')  
 price_history.pct_change()  

http://pandas.pydata.org/pandas-docs/dev/generated/pandas.DataFrame.pct_change.html

If you look below, I just need ONE thing from the numpy.ndarray object (what I get from history()). How do I pull just that one number out?

A very simple algorithm that buys and sells based on a calculated RSI value for a given stock (Relative Strength Index)

RSI = 100-100/(1+RS*) where RS* = avg(total gains)/avg(total losses)

import pandas as pd

def initialize(context):
context.security = symbol('AAPL')

def handle_data(context, data):

# Get the prices from the past 6 months (about 120 trading days)  

price_history = history(bar_count=120, frequency="1d", field = "price")  
# log.debug(price_history)  

# These variables will keep track of our total gains and losses so far within the loop  

gains = 0  
losses = 0  

# This loop will traverse the array from current day to 120 days back, 1 day at a time  
for x in range(-1,-120,-1):  
    # If current price is positive, assign it to current_gain and add it to our gains variable  
    # If its negative, do the latter  
    current_price = list(price_history.iloc[[x]])  

    print current_price  

    #price_change = current_price - price_history.iloc[[x+1]]  
    price_change=2 # temp  
    current_gain=5 #temp  
    current_loss=4 #temp  

    #print "price change: " + price_change  
    if current_price > 0:  

        #current_gain = price_change  
        gains += price_change  

    if current_price < 0:  

        #current_loss = price_change  
        losses += price_change  

    # Calculate average gain and loss thus far by dividing total gains or total losses by number of days passed  
    current_avg_gain = ((gains)/x)  
    current_avg_loss = ((losses)/x)  

    # Average overall gain/loss overall is calculated by ((prev gain or loss)*(x-1)+current gain or loss)/x  
    avg_gain_overall = ((current_avg_gain)*(x-1)+current_gain)/x  
    avg_loss_overall = ((current_avg_loss)*(x-1)+current_loss)/x  

    # We can calculate our RSI value b using the equation 100-100/(1+RS) where RS is average of x days' up closes divided  
    # by avg of x days' down closes  

    RS = avg_gain_overall/avg_loss_overall  

    RSI = 100-100/(1+RS)  

    if RSI <= 30:  
        curr_price = data[context.security].price  
        cash = context.portfolio.cash  

        num_shares = int(cash/curr_price)  
        order(context.security, +num_shares)  
        #log.info("Buying %s" % (context.security.symbol))  

    if RSI >= 70:  
        curr_price = data[context.security].price  
        cash = context.portfolio.cash  

        num_shares = int(cash/curr_price)  
        order(context.security, 0)  
        #log.info("Sold all shares of %s" % (context.security.symbol))  

record(portfolio_value=context.portfolio.portfolio_value)  

Hi John,

Which "one thing" do you need from the history call? Is it price from the last minute bar for a specific stock? If so you can do this:

price_history = history(bar_count=120, frequency="1d", field = "price")  
aapl_price = price_history[context.security][-1]  
log.info(aapl_price)  

If you need help building your algo, check out the example RSI algorithm and the pandas mini-tutorial. Hope that helps!