Hi JS,

You can log the fill price in your code to monitor the prices. For example,

prices = data[sid(24)].price  
log.info("Price for stock is %s" % (prices))  

In the backtest, your order will be filled based on the close price of the previous bar. For backtests in minute mode, this will be the stock price of the previous minute and for daily mode, it will be the previous day's close price. Note that 'price', 'close_price', and 'last_sale_price' are all equivalent in simulation.

If you run a full backtest, the Transactions tab shows the cost basis for the security. In case you're not familiar, cost basis = security price + commission costs. Your orders are filled based on the slippage model, which by default will transact up top 25% of the bar's volume. You can turn off the slippage model, adjust it to use other values, or create your own. You can also adjust the default commission settings to better fit your orders.

Cheers,
Alisa

The documentation says:

cost_basis  
Float: The volume-weighted average price paid per share in this position.  

Not sure what that means entirely, except that it accounts for multiple buys.
An example might help since I learn best from examples.
How about this as a place to start even though this is unrealistic.

2010-08-19 trade:521 INFO Buy 1 AAPL at 35.691 ttl 35 csh 4010 ptflo 4009
2010-08-20 PRINT AAPL cost_basis 35.696
2010-08-27 trade:521 INFO Buy 1 AAPL at 34.514 ttl 34 csh 3975 ptflo 4008
2010-08-30 PRINT AAPL cost_basis 35.1855

How did $35.691 become $35.696? (5 cents added)

2010-02-05 trade:521 INFO B 1 AAPL at 27.901
2010-02-08 PRINT AAPL cost_basis 27.746

Another question:
Quantopian uses Interactive Brokers for real money live trading, the lowest fees in the industry and IB has Cost Plus and Flat Rate.
Which one applies to us?
When I buy or sell 1000 shares, will each transaction cost $1 or ___?
Does cost basis reflect that, or is there specific code I can add to make it line up?
Thx

Hey Gary,
Here is an example that hopefully will help. Every bar, I log the close price. If the order has been filled, I also log the cost basis and commission. You should run it yourself and see the logs and the details in the full backtest tabs.

For JS and anyone else interested, you could simply do (cost basis - commission per share) from my code to get the price you bought the security at without commission. This would get more complicated with slippage for very large orders, but that can be dealt with by adding a few lines if needed.

For your other questions Gary, the commission structure depends how your account is set up with IB. You will pay IB commissions according to the plan you choose and if it's the Cost Plus option, the number of shares traded. IB controls all of the commission details.

Gus

Just realized something: Orders disappear from get_open_orders() as soon as they are filled completely however by keeping a list of their id's, later (next bar for example), can do get_order(id) and filled can then match the amount key value (number of shares) or some in-between number for a partial fill. A filled_price key would be handy. Does Q receive a fill price from IB when live?

The fill price is available in the UI of the live algorithm dashboard. You can also access it in your algorithm using Gus's code above

Hello Alisa,

If the fill price is available in the UI of the live algorithm dashboard, then is it computed somewhere in zipline? I'm wondering if, by importing zipline code, it could be made available to the algorithm?

Grant

Alisa, could you show how to get the fill price once a trade has been done? i.e. which object do I need to inspect? Sorry I'm aware this is an old post

Hi Thomas, no problem. Unfortunately it's still not intuitive to access the fill price. You can take Gus' code that he shared above, and run it with the commissions set to zero. That way the cost_basis variable will be derived entirely from the fill price of the security. In case you haven't seen it, here is the API documentation: https://www.quantopian.com/help#api-position

Also, Thomas, you can access the fill price when commission is not set to zero by subtracting the commission per share I define in my code — as Alisa said, not intuitive, but it works.

Hello Gus/Alisa,
Has there been any updates since, to access the fill price? Or is Gus's code still the way to go?

Thanks!

On a local installation of zipline and in daily mode I can get fill price like this:
for trans_list in context.perf_tracker.todays_performance.processed_transactions.itervalues():
for t in trans_list:
print 'transaction', t.dt, t.price, t.amount
Also I think fill price can be accessed if you set custom slippage which delegates to default implementation.

Hi Ph Quam points out that zipline provides this easily enough.

Is Quantopian going to add in the pass through for the code or is this a business decision for Quantopian to rely on UI to get actual trade information.

This is basic information needed that can be easily retrieved in zipline. Checking the price and portfolio objects can slow a fast algorithm down considerably!