algos w/ paid data sets - eligible for Q fund if only free sample data used?

Say I write an algo using one or more data sets that require payment to be able to run up to the present day, but I have not paid for them, and so my backtest duration is limited. Would the algos still be evaluated for the fund by Quantopian after 6 months, with backtests run up to the present?

As a specific example, say I wanted to use from quantopian.pipeline.filters.eventvestor import IsAnnouncedAcqTarget to filter out stocks that may be acquired (several recent Quantopian examples have done this), but don't want to pay \$85/month just to write an algo for the Q fund. Would it suffice to run the backtest over the date range 01 Feb 2007 - 10 Feb 2015 and then in 6 months, Quantopian would evaluate my algo automatically, with the EventVestor premium data? Or would my algo be excluded from the Q fund evaluation?

7 responses

It is not necessary for you to buy data for your algorithms to be considered for the hedge fund. When we do an out-of-sample evaluation of your algorithm we supply the premium data for the simulation.

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.

Thanks Dan,

Interesting. I guess I'm still a bit confused, since, for example, if I use from quantopian.pipeline.filters.eventvestor import IsAnnouncedAcqTarget I can only develop and test in the range 01 Feb 2007 - 10 Feb 2015. So, I'd be missing 1 year of data, and it would be a roll-of-the-dice to then rely on Quantopian to do the out-of-sample evaluation. I'm realizing that perhaps your primary motivation for putting out all of the various data sets is to enable your crowd of virtual quants to write unique candidate algos for the fund. However, in practice, I don't see how it can work, unless you have a lot of users paying for the data anyway, for their own purposes, and the idea is that you can give them a ring if their algos match your fund requirements (or they'll be motivated to write Q fund algos, in addition to their own).

Just trying to grasp the concept here. What am I missing?

You've got all the points correctly. We are providing data sets to help the community write more algorithms with more capacity with lower correlations to each other (using a variety of data sets can help with each of these goals). You can (and should) develop an algorithm using data sets. The people who buy the data know more about their algorithm's performance than the ones who don't. When we do out-of-sample evaluation we will use the full data set, and we will evaluate the algorithm for an allocation based on the in- and out-of-sample performance.

Perhaps your follow-on question will be, "why not make all the data free?" The economics of that don't work, unfortunately. We do what we can.

Thanks Dan,

Makes sense. Maybe I'm the only one, but I figured that there would be no point in playing around with the data sets that require payment for up-to-datedness.

I suppose that because the contest requires a backtest up to the present, one could not enter an algo with a dataset that would preclude it from running to the current day, correct? Or does the backtest triggered when the contest is entered have an override applied (as you use when you do the out-of-sample evaluations)?

You would not be able to enter the contest. Remember that the contest entry includes a live algorithm. For a lot of algorithms, you could run your paper trading live algorithm and then make trades manually yourself. (It's a pain the butt, but it's doable in a lot of cases). If you are using a premium dataset, you're reaping the benefits of that dataset without paying for it. For obvious reasons, that's not desirable.

It might be worth noting that hundreds of community members purchase data on Quantopian, and hundreds of community members trade their money through brokerage integrations. These data sets get a lot of use.

Thanks Dan,

That's kinda what I figured. I suppose your licensing allows you to pay per backtest run, or something like that, whereas you'd have no controls over individual users; they could use the data more extensively (e.g. stop/start contest algos repeatedly, to back-out information from the data sets, without paying).

I'm surprised that out of 100,000+ users, only "hundreds of community members trade their money through brokerage integrations." Are you referring to the number of folks trading real money and using paid data sets? Or do you only have a few hundred users overall trading real money?

Hundreds of people are trading through their brokerage, and hundreds are using paid data sets. There is some overlap in those groups.

People come to Quantopian for a lot of reasons. Some want to get an allocation. Some want to win the contest. Some want to trade their own money. Some people want to learn about quant finance. Some people just post a lot in the forums.