Hello Aloke,

The order method will sell/short when the number of shares is negative. I made a simple tweak to your algorithm to illustrate:

amount = -2000/curPrice # positive for buy, negative for sell/short  

The modified algorithm is attached.

Grant

Gotcha, thanks! I'm trying to write an algorithm that buys a stock when it's up 1 point from its open and sells at that day's close (very simple, I know). I feel like doing something this simple shouldn't be so difficult. Why would a limit order like this not work?

order(context.nvda, amount, data[context.nvda].close_price)  

Hello Aloke,

There's a bit of a learning curve...hang in there! What you are proposing can be done in Quantopian. You'll need to write an algorithm to run on minute data, since you'll be checking the price every minute to see if it is up 1 point from its open.

I don't have time now to work up an example...perhaps someone else can lend a hand?

Grant

Aloke, sorry you're having problems figuring this out. I'd like to get you going, and I'd like to figure out how to improve the experience so that the next newcomer can figure it out more easily.

Please do me a favor and check out the backtesting section of the FAQ. Once you've read that, I think you'll understand what's going on.

In daily mode, looking at your code:

order(context.nvda, amount, data[context.nvda].close_price)  

That is placing a limit order at today's closing price, and that order is processed tomorrow.

So, does that explain it? If not, let's figure it out.

If that does explain it, do you have any suggestions on how to make this more clear in the future? Thanks.

Dan thanks for your help, unfortunately the FAQ didn't exactly solve the problem. The main problem I'm facing is that I want to use minute tick data but I would want to execute max once a day. Since the code will be executed every minute, I don't see how I could save a global variable of the day to compare to each minute. Essentially what I'm wondering is if you can declare a variable (i.e. date using DateTime object) that can be accessed throughout multiple minute data if that makes sense?

Cheers,
Aloke

@Aloke, do you see the "context" variables in your example code?

    context.nvda = sid(6984)  
    context.max_notional = 1000000.1  
    context.min_notional = -1000000.0  

You can create a flag or datetime variable the same way.

def initialize(context):  
    context.tradedate = None

def handle_data(context, data):  
    # get current date (but not the time portion)  
    date = get_datetime().date()  
    # check trade date before continuing  
    if context.tradedate != None and context.tradedate == date:  
        # already traded today, do nothing  
        return

    # if we sent a trade order ...  
    if mysentorder == True:  
        # remember current date  
        context.tradedate = date  

Gotcha, much simpler than I thought. Thanks!