Hi all,
please help me build a mental model of howto think of the backtesting engine.
When I make a call to order a security (assuming high enough volume) can I assume the order is filled immediately?
This question only pertains when I run backtesting.
First off, very good question. "how to think about the backtesting engine?" It's important to understand the process of how trades are simulated to then understand the limitations (and power) of the Zipline backtester.
The first and foremost concept of the backtester is that it is discrete-event simulator. It 'loops' through it's code every minute. The smallest resolution for Quantopian events (and data) is 1 minute. This resolution is generally referred to in the documentation as a 'tick' or a 'bar'.
The basic steps which occur during each bar of each trading day are:
1. get orders - fetch the list of outstanding orders. These could be new orders or partially filled orders.
2. iterate through the orders and process each one - calculate any buys/sells from existing orders. How much of an order fills, and
at what price, is based upon the slippage model specified in the
initialize method (or the default if none specified). Note that the
data available to calculate fill price is only data which is
available at the beginning of this bar (ie at the close of the
previous bar).
3. execute handle_data and any scheduled functions - this is when any user defined logic is executed.
4. record new orders - done here so the timestamps reflect the current bar when entered or modified.
The above sequence is repeated each minute of the trading day.
From the above it can be seen that orders are really not 'filled' until the beginning of the following bar, or later depending upon the slippage model. Sometimes an order will get spread out over multiple bars. So, to answer the question "can I assume the order is filled immediately?", the answer is no. There is a little subtlety... even though the order logic is executed at the beginning of the following bar, the price and volume data used to determine how it is filled is from the previous bar when the order was placed. This sort of simulates an order being filled during that previous bar (sort of).
If one wants to dig into the exact order of things take a look at the class "AlgorithmSimulator" and the method "transform" . That's the main work loop of the backtest engine (https://github.com/quantopian/zipline/blob/master/zipline/gens/tradesimulation.py).
The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by Quantopian.
In addition, the material offers no opinion with respect to the suitability of any security or specific investment. No information contained herein should be regarded as a suggestion to engage in or refrain from any investment-related course of action as none of Quantopian nor any of its affiliates is undertaking to provide investment advice, act as an adviser to any plan or entity subject to the Employee Retirement Income Security Act of 1974, as amended, individual retirement account or individual retirement annuity, or give advice in a fiduciary capacity with respect to the materials presented herein. If you are an individual retirement or other investor, contact your financial advisor or other fiduciary unrelated to Quantopian about whether any given investment idea, strategy, product or service described herein may be appropriate for your circumstances. All investments involve risk, including loss of principal. Quantopian makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances.
