Live Trading Suspension - Question to Management

In thinking how I will adjust to the fact that Q has suspended its live trading functionality, I have one or two questions for management that will largely determine which path I take.

The question is whether there is any room for convergence between Q's goals and those of smaller investors. Some strategies can work for both small and larger portfolio sizes. Others, by their vary nature will only work in large portfolios with very low trading costs, while others exploit low volume stocks and could not work in a large allocation environment.

Often, it is only after testing the strategy with different portfolio allocation sizes, that that determination can be made.

From my perspective, I need to focus on strategies that make money for smaller allocations ($50k to$1 million), so that my efforts can be deployed on my real client portfolios and make them money, thereby growing my RIA practice. Of course I am hoping that one or the other allocation can be tweaked or modified to also work on larger allocation sizes, such as those pursued by Q. Anything I think has a chance to succeed would be presented to Q for consideration.

I could live with a scenario that does not allow me to trade live - no linkage to IB or RH - but does allow me to use the powerful Q environment to shape my decision-making. I would run the algo's on Q simulated trading, and create log files that tell me what positions are taken over the course of a day. I have one algo that makes decisions 3 times a day. Examining the log records just after those times would be an acceptable workaround for me - not ideal - but acceptable. I would then manually place the trades at my brokerage.

(I understand that screen scraping technologies exist that could automate that task,as some have suggested , but I think that would violate the terms of my relationship with Quantopian and possibly subject Quantopian to contract violations with their data providers. So I think Q would actively prevent that.)

So the question to Q is, would you be ok with that? That would give me an incentive to continue to maintain a very active participation and to continue to submit algo's to Q for consideration of an allocation. If so, what kind of log files would be acceptable? What kind would not?

My idea is simple: allow me to generate log files in xml format that indicate positions taken and removed. I could then copy and paste these into text and excel files that would allow me to easily make trading decisions and track the performance of the portfolio strategy over time.

Obviously this approach would only work for relatively low frequency trading strategies. But I can live with that. The question is: does this mesh with Quantopian's goals? I sincerely believe this approach would keep me and hundreds of other live traders actively involved with the Q environment. I believe live traders who put real money on the line are the most likely to develop applications that can perform well with large money allocations as they have stood the tests of reality: volume fluctuations , latency, delistings, internet glitches, etc. So Q is more likely to find its shining stars among that group.

So I'm anxious to hear from Q's management team on this issue.

4 responses

You can already do this. The log.info() or print feature already exists and the output is available from a tab within the paper trade page.

They already made their comment about this. They believe that people who write algorithms for "retail"-level live-trading are not going to contribute anything useful to their fund. I don't agree. You don't agree. Lots of people don't agree. But most of the people who don't agree very transparently exhibit their ulterior motives for not agreeing. People just want to live-trade. So it doesn't exactly help our case.

Hi Viridian.

Yes I know i can do this. The question is whether Q is ok with this or will swoop down to stop it if we formalize the procedures. I hope that they see it in their interest to support it, as we are then programming within their environment and will eventually develop strategies that work at a larger scale.

I would like their formal finding on this: what is ok, what is not ok.

I think you can safely assume going forward that type of thing will be a "not supported" use of the platform. They've been pretty clear on this point, and I don't see the point in belaboring it.

If you want to live trade there is:

Which has Python.

Q is out of the trade it yourself model through they have released http://www.zipline-live.io/ but we are yet to see if this becomes viable.