Hi,
This statarb code is based on the example in zipline. I changed the zscore method to sscore which is in the paper Statistical Arbitrage in the U.S. Equities Market. I think I replicate the paper well, but it doesn't seem to make money...
Could anyone take a look at the code and see if I made some mistake anywhere? I am very new to zipline, and can really make some stupid mistakes that I am not aware of. Thanks very much!!!
Kel
Hi Kel,
I looked at it and found differences with the paper:
- jpm and xlf sids have been entered
- ols_transform: alpha should be multiplied by 252
- place_orders: slightly modified to match the paper
- batch_transform: 61 days to get 60 returns
By the way, in the paper, there is no quantitative results about JPM/XLF just a plot of s-score! It will not be the first paper where you cannot reproduce the conclusions...
My opinion on why this strategy will not work using this algo:
Let's say an order is generated on Monday. The order will be executed on Tuesday's close price. The lag annihilates the strategy.
Hello Guy (and Yi Zhang),
Sounds correct if the algorithm is run on daily bars...orders submitted on day 0 will be filled on day 1 at the the day 1 closing price. If this "lag annihilates the strategy" what timing would be appropriate? Does it need to be run on minute bars, with orders submitted in minute 0 and filled in minute 1 at the minute 1 closing price, for example? Also, how should the algorithm handle periods when the market is closed? For example, should it submit an order at the end of day 0 to be filled at the open of day 1 (at the closing price of the the first minute of day 1)? Or should it be intra-day only?
Grant
@ Zhang, it seems interesting. Can you please share the link of the article here? I will code it again and will let you know the results.
The article link: http://www.math.nyu.edu/faculty/avellane/AvellanedaLeeStatArb071108.pdf. Thanks.
I think the strategy should be ideally executed one minute after the daily s-score is computed.
I see many alternatives: get after-hours market data, specify a time at which the batch transform is computed, force the orders to be executed on the closing price on day 0 as you wrote or run it yourself in R or Python ...
Cheers, Guy
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.
