Hey Timothy!

After you get a little experience with Python I guarantee that you'll fall in love with how simple it is and end up preferring it over spreadsheets. If you want to subtract "(H_History - L_History)" like you wrote above, just write that in Python high_low_range = H_History - L_History and if you want the mean or average of those values well why not just write that in Python too mean_range = high_low_range.mean()In fact you could boil that down to a one-liner mean_range = (H_History - L_History).mean()

Overall...

H_History = history(context.days, "1d", "high")  
L_History = history(context.days, "1d", "low") 

mean_range = (H_History - L_History).mean()  

Print that result and it should be a collection of the average range for each security.

Thanks. That helps a lot. But there is a new challenge. I am wanting the PREVIOUS days' data, not including today's data. Frankly I am a bit surprised there that 'history' cannot start 'x' bars back and go for 'y' bars back from there. Or at a minimum, the ability to exclude the current bar and use the true "history" (ie data from the past but not the present).

I tried simply subtracting history(1, "1d", "high") but that doesn't work. I tried moving the history(context.days*+1**, "1d", "high")* to the 'before_trading_start(context)' section, but that returned empty values for the highs. I need to simply remove the last line from history(context.days*+1**, "1d", "high")* . Or find the mean of the first 5 lines of data, but ignore the 6th.

:-/

Hmmmm .... I jsut found this in the help page:

def handle_data(context, data):  
    price_history = history(bar_count=2, frequency='1d', field='price')  
    for s in data:  
        prev_bar = price_history[s][-2]  
        curr_bar = price_history[s][-1]  
        if curr_bar > prev_bar:  
            order(s, 20)  

I think I can steal and modify that to fit my needs.

So yes ignoring the latest bar of data is a very common occurrence when doing time series analysis, but even more so it is a common occurrence in programming in general so Python has this super powerful operation called slicing. Here is a good Stack Overflow explanation of slicing. The object that history() returns allows slicing as well so if you want everything except the last value you can just say [:-1]

price_history = history(10, "1d", "price")  
history_except_today = price_history[:-1]  

Check out my response in this thread for some good places to start when learning to use Quantopian or learning Python. It really is worth struggling through things like this as the struggle will make you a better programmer which will in turn make you a better algorithm writer.

J

Is there also an efficient way to identify any securities that have had a sudden drop of, say, 50% of the range between max and min?

@garyha

That's kind of an arbitrary question, a drop in relation to what price? The max and min over what time span? These kind of questions are similar to how you bound the definition for max drawdown.

Any. Easiest would be fine. Dev choice.