So glad I had the TV on for a brief moment. This is the best news I've heard today! Coded many algos at CQG for 5 years, helped others for 2 years after and now trying to build a media production empire. Damn I only wish there were more hours in a week... Hope to post something soon.

Very exciting news. Congratulations Q.

What happens if the fund is under water? Do algos who generated profit still get paid?

Very exciting news, congrats and nice work Quantopian!

@all thank you

@Miles Please check out the allocations page. Quoting that page, "You will be paid a share of your algorithm's positive returns, regardless of other algorithm writers' performance."

@Quantopian I just read the "Quantopian Terms of Use" . Based on what I read it's my understanding before a financial allocation would be made to an author's algorithm, a separate agreement ("Supplemental Agreement") would need to be approved by both Quantopian and the author. Is my understanding correct? And if so, would you kindly provide details on how I might obtain a copy of the "Supplemental Agreement", so that I might review it as soon as possible?

Hello Michael - Yes, you are correct that before we make an allocation to an author's algorithm, we need to reach an agreement with the author. We often refer to it as a royalty agreement. The broad terms of the agreement are covered on the allocations page. We haven't made the document publicly available, however. I will look into creating a summary that we can share more broadly.

While I, so far, have not been in a position to put up an algorithm for your consideration, I greatly appreciate what you are doing with this platform. I believe it also enables myself and others to pursue a wide range of goals using your platform. For that and your other contributions I thank you! May this turn out to be "win / win" for many!

@Dan Dunn I've read over the allocations page already. I'll hang tight and look forward to the royalty agreement being made available. Thanks for the quick reply, especially on a Saturday!!

Congrats Fawce and team: awesome marker in history. I tend to write algo's that you guys dont like, but then again, I might try to make them more hedgefund compliant now we can earn serious allocations

Congratulations!

A few questions:

• Back-of-the-envelope, to allocate all of the Point72 committed $250M, how many community members would you expect to get allocations (presumably, you'd need diversified capacity up to 6X$250M = $1.5B, if leverage is included)?
  
• What mix of equities versus futures do you expect for the Q fund?
  
• What is the path to get to 1M users and $10B of assets under management? Any vision for how to scale?
  

Congratulations Quantopian!

It would be interesting to know how you selected the algorithms: from the contest? algorithms that were in paper trading? users spontaneously submitted those algorithms for your review?

I'd also love to see the pyfolio analysis of those algorithms with out of sample performance and to hear your reasoning behind the selection of those algos.

Quoting Delaney Granizo-Mackenzie "I would try to stick to the Q1500, into which we put a lot of work. It's the universe required for algorithms that receive an allocation from -- and are traded by -- Quantopian.". Does that mean all the selected algorithms make use of Q1500? That is not surprising but given the actual resources limitation it's quite challenging to use that universe in many cpu/memory intensive algorithms.

It would also be interesting to hear the extent to which the algos conform to the type described on https://blog.quantopian.com/a-professional-quant-equity-workflow/ or if they are more single-factor, few stock, retail-style algos? Breakdown of equities vs. ETFs? Etc. Any nitty-gritty details that could be shared would be of interest.

Also, have any of the folks who got funding decided to share their code with Quantopian?

@Luca - The algorithms that were selected came from all over. Some of them were entered in the contest, some of them were pointed out to us by excited authors, and most of them were identified by our evaluation process. The automated evaluation process screens almost every algorithm written on the Quantopian platform, and we're continuing to refine that automation to make sure it's casting a wide enough net. We don't want to overlook any good algorithms, for obvious reasons.

As for the pyfolio analysis, we're not currently planning on releasing that level of specific information on the funded algorithms. But, if you watch the "How to Get an Allocation" webinar you can hear VP of Quant Finance Jess Stauth do detailed analysis of what she's looking for when she's evaluating algorithms for Quantopian’s use. She is using the criteria we list on the allocation page, plus more color. It's not exactly what you asked for, but I think it's just as informative.

On the question of the Q1500US: almost all of the algorithms trade in equities found in the Q1500US. There are plenty of viable Q1500US strategies that run on the platform today. That said, we look forward to increasing the platform's power so that we can support even more Q1500US strategies.

@all - I have a couple other thoughts to share. First, the questions about ETF v. non-ETFs, or about the future relative size of the futures portfolio, or other nitty-gritty details are generally ones that we're not able to share. If we talk about our fund or its portfolio, that might be viewed as a form of public solicitation, which we cannot engage in. In our discussions we carefully stick to the question "What are the criteria we are looking for to make an allocation?" and we eschew the questions that sound like "What does your fund or investment decisions look like?"

Also, our recent work has driven refinements in our algorithm review process, and we're working on getting that feedback into the product and into our education. I can share a few bullet points, though each of these points needs a lot more than I'm ready to say here:

• We need to tighten the window of desired beta exposure of each algorithm. The contest guidelines were to keep the rolling beta to SPY between -.3 and +.3, but the algorithms we selected were in an even narrower window. We're also expanding the way we measure an algorithm’s risk beyond looking at a trailing 6 or 12 month beta-to-SPY, to include a forward looking beta exposure for each algorithm that is derived from summing up the beta exposures of the algorithm’s individual holdings. We'll be publishing better guidelines and how-tos on managing beta going forward.
  
• Too many algorithms have a net long exposure, even when their beta was low. We're going to revise our criteria to ask for not just low beta, but to ask for net-zero dollar exposure as well.
  
• Quite a few algorithms are running into problems with sector exposure. An individual algorithm might have good, low beta, no net exposure, and still have a risky exposure to one or more sectors. We will give more guidance and advice on this topic, too.
  
• There are other risks that people continue to overlook that we need to make easier to measure and manage. For example, there are too many algorithms that rely on returns generated from investing in small and microcap stocks.
  
• We still see a lot of algorithms that trade in too few names, resulting in high single-stock exposure.

Congratulations Quantopian,

You're a dream come true and the best is yet to come!