If I understand your question correctly, cant you keep track of max and min values?
For example,

handle_data:  
    # Check that this part runs prior to Q1 earnings  
    if data[stock].price > context.max_price_for_quarter:  
        context.max_price_for_quarter = data[stock].price  
    if data[stock].price < context.min_price_for_quarter:  
        context.min_price_for_quarter = data[stock].price


    # Now run this after earnings by checking the date/month  
    if data[stock].price < context.min_price_for_quarter:  
        buy_stock()  
    # and vice versa for sell  

Another bit of practice in python...

I've built a number of S&R type indicators/pattern detectors over the years, never in python though, not very easy that. But here's a variation. The below code looks for isolated lows which then get added to a list for each security. In the backtest you'll see that it looks back 200 periods (everyday) and identifies lows that are the lowest low to 30 periods to either side. A significant low. It rounds the price and adds that low to the support list.

def BuildSupportLevels(context, data):  
    result = {}  
    l = IsolationPeriods  
    lowAll = history(LookbackPeriods, "1d", "low")  
    for stock in data:  
        result[stock] = []  
        lows = lowAll[stock]  
        try:  
            for i in xrange(IsolationPeriods + 1, LookbackPeriods - (IsolationPeriods + 1)):  
                if (lows[i] < min(lows[i+1:i+l])):  
                    if (lows[i] < min(lows[(i-l):i])):  
                        #print("Lowest low {0} {1:.2f}".format(lows.index[i], lows[i]))  
                        level = round(lows[i], 0)  
                        if (level not in result[stock]):  
                            result[stock].append(level)  
        except:  
            continue  
    return result  

In the trading test it looks for a rounded current price that matches any of the lowest lows (and highest highs for resistance). If so it buys them.
The sell short code is commented out. I may readdress this code in the future and fix issues I as I mull over the technique. There are lots of holes in this approach but I'm not gonna get into them right now.

    SupportLevels = BuildSupportLevels(context, data)  
    for stock in data:  
        if (round(data[stock].close_price, 0) in SupportLevels[stock] and data[stock].mavg(7) > data[stock].close_price):  
            eligibleLong.append(stock)  

Thanks Ajay. The quarter min won't really work, but I think you are on the right track.

Market Tech, thanks for the response. I'll see if I can incorporate some of that code into my algorithm and let you know how it works.