Back to Community
Backtesting to 6 months of live trading - a tearsheet analysis

$10k initial capital

Updated with Dr. Wiecki's suggestions

Loading notebook preview...
9 responses

Hi Gene,

Thanks for posting, this is exactly the type of sharing I was hoping for when writing pyfolio. Your tear sheet would be more informative if you ran the backtest over a longer period up till yesterday and then, when calling create_full_tearsheet() you set live_start_date='2015-1-1' or whenever you launched this to paper-trading. That will make clear how it performed since then. If you're worried about IP, you can also set hide_positions=True so that no position names will be displayed.

In any case, sharing tear-sheets without the algo code is a great way to make us aware of a strategy you think would be interesting for the fund.

Thomas

Disclaimer

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.

Updated with Thomas's suggestions. Also ran a Bayesian tearsheet.

Gene, thanks for the update.

It seems that your OOS performance is more volatile than the in-sample. The Bayesian analysis puts some statistical rigor behind that (row 4) where you can see that the difference in volatility between IS and OOS is statistically significant. In any case, the strategy still looks appealing.

Thanks for your comments Thomas. I might share a tearsheet analysis of one of my first algo that has 9 months OOS just for fun. Is that something you would like us to do in general?

Also, I had to use "hide_positions=True" instead of "hide_pos=True".

Charles: Yes, definitely, that'd be fun. It's also a great way to get feedback on an algorithm and ideas for improvements. Justin is great with that.

I fixed the hide_positions kwarg in the post above.

So here we are 12 months of out of sample later. I think this is a good reminder that live trading is the best way to validate your trading strategy.

Loading notebook preview...

Thanks for posting these updates, it's a refreshing dose of the humbling reality.

Does the curve represent live trading data with real money?

It would be really interested to see the backtest for that period with the same algo and live data compared so it could be seen how accurate the backtest modeling is. Is it possible to do something like this?

The curve above is a backtest, thats how the bayesian stats are generated, but it has actually been paper trading since the indicated start date. Thankfully this is not real money. I haven't gone through the transaction logs to see how the two compare but here is the above notebook with the live results fetched with the get_live_results function in research. It stops in April because that is when Q2.0 launched and it broke the algo so i had to stop it and restart it.

Loading notebook preview...