Late last year, USV invested in Numerai, a hedge fund that uses data scientists all around the world to “crowdsource” stock price predictions. I blogged a bit about Numerai then.

If that business model wasn’t cutting edge enough for you, the Numerai team has now gone a step further and issued a crypto-token called Numeraire to incent these data scientists to work together to build the best models instead of just competing with each other.

When I read the Numerai blog post about Numeraire yesterday, I tweeted this out:

This is all pretty out there stuff in a world, hedge funds, that has more or less done things a certain way for the last thirty years. I’m not saying hedge funds haven’t innovated, they certainly have, but I don’t think anyone has attempted to change the behavioral economics that underpin hedge funds in quite the same way that Numerai has. It is, if nothing else, a fascinating experiment that will tell us a lot about crypto-tokens, machine learning, and behavioral science.

I must admit that some of this is over my head. I’ve read the Numerai blog post as well as the Forbes and Wired posts several times now and I am not sure if I could explain all of this perfectly at a dinner party. But I am super excited that USV has invested in this audacious experiment and I look forward to seeing how it all pans out.

#crypto#hacking finance#stocks

Comments (Archived):

  1. aminTorres

    “I am not sure if I could explain all of this perfectly at a dinner party.” even if you could, don’t ;)kidding aside, their visual aesthetic is really interesting. Somehow it works. I did not think you could make a financial services website/product/video look artistic.

  2. JoeK

    Perusing through the blog post as well as whitepaper leaves me with the impression that this is simply an opaque lottery, in which the participants have no way of verifying whether Numerai pays everyone correctly. In the end, how different is this from gambling? Beyond that, the opportunities for profiting from hedging and arbitrage (by betting both ways) seem boundless.

    1. fredwilson

      i think its the exact opposite because all of this is coded in a smart contract that pays out automatically without human intervention

      1. LE

        I think his point is ‘but how do I actually know that is happening’.

        1. JoeK

          Exactly. They state that they pay out if predictions are met, but Numerai keeps the real model/trade/transaction data secret, so there is no way of verifying that the process is fair.

          1. karen_e

            These conversations always come back to trust.

          2. JoeK

            No one ever seems to admit that.

          3. Amar

            And the answer seems to be “trust the algorithm” 🙂 riiiight

        2. aminTorres

          The smart contract has logic that is defined by the parties involved in the interaction and always act as scripted. So it always makes a unbiased decision when the conditions in the contract are met.Gambling is not possible.If Numerai works in alignment to the blockchain, the contract itself pays and not Numerai.

      2. Twain Twain

        Re without human intervention, here’s Alan Turing: “One must therefore not expect a machine to do a very great deal of building up of instruction tables on its own. No man adds very much to the body of knowledge, why should we expect more of a machine? Putting the same point differently, the machine must be allowed to have contact with human beings in order that it may adapt itself to their standards.”Humans (not just lawyers or bankers or engineer-miners) need to have the calibration tools to tune the systems according to our human standards — not just because it fits the “code is law and should be automated” principles espoused by the Satoshis and his disciples.I’m in favor of homomorphic encryption. It’s some of the aspects of Deep Learning and its opaqueness for audit trails that may become issues.It’s especially paradoxical because Blockchain’s core mission is to make data more transparent and accessible to more people rather than in the hands of governments, the banks and large corporations.

        1. Lawrence Brass

          Fortunately there is an auto-limiting factor on the ‘weirdness’ of automation because value is created or realized at the exact moment of interaction with humans, not anywhere else.So I propose “without human intervention or involvement there is no value” created or transacted. In this craze of excluding humans from every activity and automated trust, I would like to remind ourselves that what makes the economies tick are us humans, and not our machines.

      3. Twain Twain

        Also, here’s the connection between why it’s VITAL to solve Natural Language Understanding in AI and how the blockchain for finance works … https://uploads.disquscdn.c…The lead inventor of IBM Watson that beat ‘Jeopardy’ and Bridgewater’s AI just founded a new AI company to try to get to “General Intelligence AI” that can support corporate decision-making.The core piece of the puzzle for everyone (the “Holy Grail”) is … the Natural Language Understanding piece.We can’t get there with the Scientific Method inherited from C17th Enlightenment.

  3. LE

    I must admit that some of this is over my head. I’ve read the Numerai blog post as well as the Forbes and Wired posts several times now and I am not sure if I could explain all of this perfectly at a dinner party.That challenge alone makes this worthwhile to try to read and understand.Question though, isn’t some form of the following already been modeled in computer games? It seems obvious so my question is if that is the case then what have the results been?Imagine the prisoner’s dilemma in a world that exists entirely on a blockchain. Now suppose the prisoners are issued a cryptocurrency similar to a normal money except for one small change: it is programmed to self-destruct whenever anyone goes to prison. By defining the money in this way, the prisoners’ fates are now financially bound. Prisoners in this scenario realize that if they don’t keep the other prisoner out of jail, they will lose all of their money with certainty.

    1. JoeK

      In fact, it seems the complete opposite to me. In that scenario, you would want people to go to jail, because the remaining currency, in which you hold a stake, becomes more valuable given that it becomes more scarce. I don’t get the logic in their statement.

      1. Mikesonds

        All of the money will be gone, not just some of it.

    2. creative group

      LE:just when we thought we understood the premise you opened this paradoxical paradigm.

  4. LE

    I think this could be really really big. Because it involves at the core psychological manipulation and (not saying this in a negative way) taking advantage of free labor and the nerds needs to be recognized and agreeing to play with no immediate reward tied to time invested. That said I don’t think the article clearly detailed the exact benefit to helping your fellow man at least not in a convincing way that was obvious. In other words by recruiting other scientists to improve the model is the only benefit that you don’t loose what you have already? Or is the potential payout (other than no loss) far greater if you get 1000 others to help?

  5. Richard

    Hedge Funds are a non productive tax on investors, particularly retirement accounts. Why would usv support this?

    1. fredwilson

      if this works it could massively reduce the cost of producing high performing investments for people. that’s a good thing.

      1. Kevin__Fung
      2. JamesHRH

        Markets are supposed to allocate capital to productive businesses, kind of like USV does.Focusing on the ‘cost of high performing investments’ side of the coin means that the tail wagging the dog.Does Numeral care where the capital is allocated? If not, its a tax on the market.

    2. Lucas Dailey

      This is absolutely productive. The amount of R&D that will be done to better understand pricing and economic trends will be considerable. In fact it grants a business model to areas of machine learning that might otherwise remain academic. And of course does it in a meritocratic way that anyone can participate in.

  6. josephcohen

    How does one invest in the fund?

    1. fredwilson

      it is closed for now and funded almost entirely by the founders to date. but that will change soon.

      1. josephcohen

        interesting. seems like a core part of the idea?

    1. jason wright

      what’s this? is there a translation available? thx.

  7. jason wright

    Is this legal?

    1. Lawrence Brass

      street legal

      1. jason wright

        that would be Street legal?

        1. Lawrence Brass

          Hang a flashlight in the front, red tail lights at the back and… you are ready to go. 🙂

  8. pointsnfigures

    From their blogpost, “The stock market presents a situation similar to the prisoner’s dilemma. The market would be better off if market participants collaborated, but rationally they don’t.” I am not sure this assertion is correct-and I am not sure there isn’t already cooperation.1. If I can crush my adversaries/competitors, all the surplus goes to me. (Monopoly)2. There is cooperation via liquidity, and transparent price. The underlying economic incentives of each participant manifest themselves in the bid/ask spread. 3. There is certainly behind the scenes cooperation-for example many option and derivative markets are “call around markets”. 4. In many cases, liquidity providers will cooperate in order to keep the market moving. Markets are a continuous game, not a single event game. This is why Prisoner’s dilemma might not be a good way to analyze the market.That being said, what they are doing is kind of cool. We will have to see how it all works out.

    1. Girish Mehta

      Naval Ravikant here -“…a lot of virtue comes from the “iterated” Prisoner’s Dilemma”.

      1. JamesHRH

        “All the returns in life come from compound interest.”The longer and stronger your relationships, the more successful you will be.

        1. Lawrence Brass

          There is a virtuous relationship between long term value and trust.I think that immediate value can be hacked, but long term value cannot.

  9. JohnLaGrou

    We’ve lost our moorings. Value is created by delivering value to others, and investing once reflected that ethic. Today, it’s about microsecond timing differential trading windows and layers of opaque financial derivatives, all held together by historically high debt ratios. A casino mentality. Our 60% middle-class real-wealth and income-growth proportionality continues to fall, as the top continues to accelerate. My friend, this is unsustainable. It’s a house of cards built on a sandy foundation.

    1. kenberger

      …and we now even have Frank Underwood in the white house, speaking of house of cards.

    2. Lucas Dailey

      Sure, but that’s our financial system as (poorly) defined by our policies, laws, and regulations. A tool like this at least helps move toward more democratic (and socialist?) means of profiting from that.It’s easy to imagine Numerai/future-competitors that opens up investment to the general public. From there it could be used as a means of getting public investment a better slice of the pie. S&P index funds will be dinosaurs eventually.Yes, there are structural problems with our financial system and economy. This could lead to some ameliorations of those problems, but at worst it’s doing no harm. We all have a moralist urge to holistically fix any broken system we touch, but really just trying to do one small piece well is hard enough. No one should feel bad about that.

      1. JohnLaGrou

        “…that’s our financial system as (poorly) defined by our policies, laws, and regulations.”Yes, sadly. Which is really the only fiscal conversation that matters, to those interested in long-term, all-class socioeconomic health. Otherwise, brilliant investment tools like Numerai may become little more than fancier deck chairs on the Titanic.

  10. Pete Griffiths

    Fred”… the behavioral economics that underpin hedge funds…”I am curious what you mean by ‘behavioral economics’ in this context.

  11. William Mougayar

    The other thing that’s new about this is that it’s the first non-crypto business that created a token. I predict there will be many more like that this year. And their rate of success may be higher than virgin crypto businesses because these start with a sound business base, unlike blockchain startups who have a lot to prove.

  12. jason wright

    So this is built on the Ethereum blockchain, but incentive rewards are paid in bitcoin. Is that right?

  13. asg749d

    Reminds me of Estimize. They have a similar concept.

  14. howardlindzon

    I am so glad you are saying that some of this is over your head. Ryan Delk introduced me to Richard in 2015 to invest and I really could not understand it and so I had to pass. I am starting to get it but still don’t. I invited Richard to speak at Stocktoberfest though.

  15. Tom Labus

    Someone will try to game this. It’s the nature of markets and those who play in them.We’ll see.Stanley Druckenmiller taught his wife a trading system but only enough to react to the trading system’s alerts. Moves were made solely on those alerts. No macro info and gut. Did ok for awhile.