Back to Community
Any walk-forward framework...


Has anyone have implemented walk forward optimization in Quantopian.

I find it difficult since the standard backtest feature is not callable during
live trading.

3 responses

Hi Yagnesh,

I've spent a fair amount of time thinking about this (see e.g. and While the first one discusses walk-forward optimization, it does not implement it and unfortunately it is not straight forward to do with zipline/Quantopian for the reason you mention. What you want to do is freeze the algorithm (perhaps by pickling it) at a certain point in time and then play it forward with different parameter combinations.

Having said that, we are currently exploring more clever optimization techniques with zipline (but not walk-forward yet).


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.

If you can frame your model in vectorized terms (for instance, in Research to begin with), it's possible to optimize your parameters over history data. Optimizing actual event-driven sub-backtests I would think would be much too slow to do in 50 seconds.

I was under the impression that the API announced on would allow for more extensive number crunching (e.g. churning over large baskets of stocks either in the background during the trading day or overnight). If so, would it also allow for optimization routines that take longer than 50 seconds?