@beginer,

On https://www.quantopian.com/posts/quantopian-managers-program-algorithm-selection-and-compensation, Fawce gives a sense for the vision (posted Feb 13, 2015):

"I think it is critical that we be able to evaluate a huge number of algorithms without code access. That's a capability that takes us to an unprecedented level of scale for research. I think that's really important, since we all believe that talent is scattered at random around the world. We won't realize our vision of finding the very best talent without a very high scale solution to our initial evaluation of the algos. Ethically, I like that we are not asking for code access when we aren't compensating the quant.

Success for the fund means scaling to 10s of billions in assets. At that level, we will be compensating the quant in a very significant manner and we will have fiduciary responsibility for our investors. The ethics and practicalities of that situation are totally different from where we are now. I expect that we will have a different kind of relationship with the quants in the fund. Vetting quants will use all the automated techniques we're creating for the contests, plus new qualitative and person-to-person checks we will create in the future. There is a lot of work to do. For now, we're focusing on the most scaleable and fully automated things we can do to evaluate the huge number of algorithms already on the platform."

My read is that when they start to try to pull in institutional money, Quantopian will need to "smell" an awful lot like every other hedge fund (the "fiduciary responsibility for our investors" part). No more black-box algos by anonymous authors from the far reaches of the globe. But that's just my 2 cents. So, if you are aiming to eventually get a quant/fund manager job at Quantopian, my hunch is that there will be something like it down the road. Good luck!

Thanks Grant. That explains a lot. Having worked in several financial institutions I have an understanding of the rules and regulations that govern such firms. I agree it won't be possible to remunerate black-box algorithms by anonymous authors from far reaches of the globe given the anti money laundering, anti terrorist financing etc. etc. laws and regulations. Have to see how things shape up in the future.

If they build their own internal trading teams, I wonder what will happen to all the intellectual property generated by the crowd?

"With Quantopian's vision to become a crowd sourced hedge fund what are the possibilities for some one in crowd to become a full time paid manager with Quantopian? Is that feasible? Or is this only for people building part-time algos on side along with their other careers? Secondly do they have plans of building an internal team to build and develop trading strategies besides those from the crowd? Would like more clarity on how they see things working out within 1-2 years time frame."

Hi beginner,

Great questions. Quantopian is not and will never build an internal team of quants to compete with the quants in our community - we are strong believers the power of properly aligned incentives. We do have a small team of researchers and data scientists, myself included, working on designing the algorithm selection and portfolio construction process that will drive the Quantopian fund. It's an exciting project and we're very happy with the progress so far.

In terms of what you could expect if your algorithm is selected for inclusion in our fund, that's going to be driven by two numbers, the capital allocation and the returns it generates. At launch we'll be a small fund, and will make small allocations to each algorithm, probably on the order of $500k-$1mm per algo. Regardless of the fee split and returns generated, that level of allocation will probably mean it's a sideline for most folks at first. As the fund grows and as we validate the capacity of the algos we source from 'the crowd' we plan to ratchet up the allocations we make to each algo. In the 1-2 years time frame I don't think its unreasonable to expect that our best algo writers could be earning enough in fees to make Quantopian their sole source of income -- though we won't require them to do so.

In terms of the legal details our approach is going to be modeled on a revenue share limited term software license. Algo authors will license their code to Quantopian and Quantopian will pay the author a share of the returns generated, the author continues to own their own IP in this model. We expect to be able to license algos from authors both inside and outside of the US at launch.

Hope that's helpful - please feel free to follow up with more questions here or directly to [email protected]
Jess

Hi Jess,

This is helpful, do you have a timeline that can be shared that would show:

1. When will futures, options and forex become available?
   
2. When will you start trading investor money?
   
3. Will you start trading investor money before futures, options and forex are available?
   
4. How many algos will you need in order to accept AUM?

When, how do you see this playing out?

Thank you,

Hi Mike,

Our timeline is basically to move as fast as we can, while being thorough enough to develop high confidence in the component algorithms we fund.

"When will futures, options and forex become available?" We're working on futures now, so I feel comfortable saying that it will be available later this year. We haven't kicked off projects on options or forex, so I think its unlikely we'd get to and complete either of those this year - but that's about as specific as I feel comfortable forecasting.

"When will you start trading investor money?" Our current target is 2016, but of course this depends on several key steps between now and then.

"Will you start trading investor money before futures, options and forex are available?" Yes. Our current plan is to launch with US-listed Equities & ETFs.

"How many algos will you need in order to accept AUM?" The right number of algos depends on the capacity per algo and the projected 'decay' of signal from any given algo. My back of the envelope target is to have 20-40 algos in hand with good confidence that we've tapped into a renewable source for quality algos going forward.

"When, how do you see this playing out?" That's a bit ambiguous, but at a high level our plan is to launch with investor capital in 2016 and start building track record. A solid track record will help us attract more investor capital and allow us to increase the allocations we make to each algorithm within its capacity limits, as well as allocating to additional strategies. Along the way we plan to round out the platform product offering so that we cover all asset classes and tradable markets, which will open up the largest playing field for our community to build profitable algos with low correlation to each other and to common indices.

Best, Jess
edit: meant to include a link to the /fund page which has more detail on what we are looking for, and not looking for, in the fund.

Hi Jess,

What's the story on legal matters, compliance, etc.? This has come up a number of times on the forum, and I don't recall any clear answers. Perhaps you'll be able to get some seed money from your VC's to fund the initial 20-40 algos, but beyond that, might it be a challenge? It just seems like you are looking to go to traditional institutional markets for hedge fund products, that will be regulated. Each institution will have a set of unique requirements, as well (I'm guessing none of them include provisions for black-box, globally crowd-sourced algorithmic hedge funds). Or maybe the approach is "We'll cross that bridge when we come to it."

Grant

Hi Grant,

Legal matters are progressing nicely. There are no blockers that we've encountered yet, it's just work that needs to be done. As for institutional diligence requirements, I'm not too worried about that either. Quantopian is actually in a position to be a lot more transparent to them than most quant hedge funds. With our open-sourced backtester etc. we're a lot more of a clear box than most.

Best,
Jess

Hello Jess,

Thanks for the update. I'm not sure how the open-sourced backtester fits into the discussion, but oh well. I guess once you get a track record of a few years with 20-40 managers, it'll start to take shape. There is a Wall Street precedent for dreaming up and legalizing innovative financial products (the recent mortgage crisis comes to mind). So, I can imagine there will be room for the world's first crowd-sourced hedge fund in the mix.

Grant

Jess,

A couple quick thoughts:

• It would be best to hear from your lawyer(s) directly. Could they respond here? Or maybe if they've written opinions, you could post them as PDF's?
  
• How many institutional investors have you talked with? Which ones? What have they told you, specifically, regarding your crowd-sourced hedge fund idea?

Grant

@Grant, are you sure you're not an investigative reporter in disguise? Are you working on a secret 60 Minutes segment?

@Market Tech, just curious really, and skeptical. "Legal matters are progressing nicely." doesn't really say much, but if a lawyer has written something up, and it could be posted, then it would be substantive. Just seems kinda odd that it would be o.k. to have a portfolio of black-box algos, but I guess that as Jess points out, the details of hedge funds are not published, either.