Back to Community
Intraday algorithm using gaussian copula [deleted post]

Here is an improved version. It is an intraday trading strategy. No overnight positions.

7 responses

Thanks for sharing.

When you create tweaked versions like these, how are you protecting yourself from overfitting? It feels like you're well down the slippery slope.


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.

Fair point Dan. What do you suggest? Where do I draw the line between parameter optimization and overfitting?

What does the code do, in words/pictures/pseudo-code? A starting point in considering over-fitting is to ask why the code would be expected to work in the first place.

Where do I draw the line between parameter optimization and overfitting?

It was helpful for me when I saw Fawce write to the effect that it is a constant battle, and I took Dan Dunn's advice to heart to hold back some data, doing that with structural changes and some inevitable tweaking of parameters at first only thru 2015 or 2016, then when confident, going beyond that. Always eye-opening.

Nice work @Rooster!

@Dunn, "Overfitting" is certainly a concern, yet I don't intrinsically see it in @Rooster's code.

Every once in a while I read
which I interpret as telling me to do more conscious testing wrt in-sample/out-of-sample data, along with monte-carlo-like tests for process invariants.
This all takes a lot of time and effort and sometimes feels like "analysis-paralysis", which I'm personally prone to.

Getting back to this algorithm, the issue I'm noticing in algorithmic trading is the "time-regime" problem, whereby methods work best for certain market conditions during certain time frames. Perhaps it's a dream, yet it would be nice to have an analytic on/off signal for when to trade(or not trade) this strategy...the copula version of when a pair/portfolio of assets is co-integrated so that mean-reversion works and produces a profit!


Aqua -

Why did you remove the code and your description of it?


I found a way to actually trade it and hence everything changed.