Hedgefund Startup Powered By Crowdsourced Code

In the financial sector, everyone is looking for a new way to get ahead. Since the invention of the personal computer, and perhaps even before, large financial institutions have been using software to guide all manner of investment decisions. The turn of the century saw the rise of High Frequency Trading, or HFT, in which highly optimized bots make millions of split-second  transactions a day.

Recently, [Wired] reported on Numerai — a hedge fund founded on big data and crowdsourcing principles. The basic premise is thus — Numerai takes its transaction data, encrypts it in a manner that hides its true nature from competitors but remains computable, and shares it with anyone who cares to look. Data scientists then crunch the numbers and suggest potential trading algorithms, and those whose algorithms succeed are rewarded with cold, hard Bitcoin.

The encryption, which is reported to be similar to homomorphic encryption, is necessary to protect Numerai’s proprietary secrets. While it slows down computation, it does allow the company to more comfortably farm out its algorithm development to the wider public. Machine learning algorithms are used to determine which trades to make, and under which conditions. This isn’t the first time machine learning has been used in the financial sector, but it is interesting to see it applied in a crowdsourcing context. This allows the company to pick the best algorithms submitted by a wide base of coders, rather than relying on a handful of talented hires. Numerai is then perhaps most interesting for its hiring strategy, rather than doing anything particularly groundbreaking in a technical sense.

The fund has been trading for a year, and the founder, [Richard Craib], claims it is turning a profit. Investors are flocking to the company, which has recently completed a multi-million dollar funding round. The article sums things up rather well, however:

The company comes across as some sort of Silicon Valley gag: a tiny startup that seeks to reinvent the financial industry through artificial intelligence, encryption, crowdsourcing, and bitcoin. All that’s missing is the virtual reality.

Like it or not, machine learning is here to stay. If, like this writer, you could use some help getting to grips with it all, check out our machine learning primer.

17 thoughts on “Hedgefund Startup Powered By Crowdsourced Code

  1. While interesting, I fail to see how an ever growing number of machines that trade based on statistical analysis is good for anyone but the people who own them. They don’t trade using knowledge, they trade purely on numbers. They don’t support companies, they just own part a small chunk of them for a short periods of time. Honestly, I think the market and the companies it purportedly supports would be much better off if there was a percentage fee for all trades that don’t last at least one year.

    1. It is like Superman III or Office Space where the HFTs get to collect all of the fractions of a cent on all transactions, but not quite. They are inserting themselves into auctions where they trick the buyers and sellers out of these small amounts by intercepting the buy and sell bids before they are answered and pretty much high speed bid snipe and then resell to the MIM’ed buyer, all day every day rarely taking a loss because they can rely on a massive ping time advantage in their currently legal man-in-the-middle scam. Free money…
      But when the big players all do the Bernie Madoff career shuffle between government regulation and big trading houses of course the system will always favor the most profitable crooks, those are the crooks who hire your chosen for you elected politicians and who write the laws you have to follow.

  2. Fat-fingered the report flag on the previous comment. Sorry.
    High frequency trading is a cash cow for the connected companies at Wall Street. It is also a recipe for a market crash. Hedge funds like the one in this post are positioned to take advantage of market trends and lags in trading but that responsiveness can be a self fulfilling prophecy and tank the market.
    HFT should not be legal.

  3. Yeah, I’m of two minds on this as well.

    Interesting paradigm (there, I said it!) for outsourcing the work, but like most else happening in the financial world, this is just something else that sucks out value from anything, including from people, and rewards it to only a few.

    Would love to see some open finance to go along with our open source. Might be time for Basic Personal Income, if we economically survive the next 4 years.

  4. The actual news here is the use of homomorphic encryption (or at least a reduced analogue to it).
    *This* is the technology that will change everything, not a new method of playing the microsecond money game.
    When I can keep my data encrypted *during calculation*, as well as at rest and in flight, then running my own services becomes unnecessary. Granted the method used here is very limited, and fully homomorphic encryption is as yet impractically slow, but watch this space.

    1. I really believe that they’re stretching this feature. I think they’re providing something of the homomorphic concept – alter the base data to be unrecognizable from the source, share it, modify it, and be able to revert it. But for source data like company share prices that seems fairly simple. But it’s not the full-fledged homomorphic encryption being worked on by researchers.

  5. Trading is a game of speed, of mili seconds. Trading software is optimzed to trade absolutely as fast as computerly possible. Trading companies pay millions of dollars to have their trading servers mere feet closer to the wall street servers. The time saved by the signal running a few feet less of the cable puts your company at a great advantage over the other companies. They are encrypting their code and beating anyone? Could be they are gathering crowdsource code, paying the contributors a penance. VC capital is pouring in, they live like kings. Company folds, who cares. Golden parachutes anyone? They take their code to the big boys, who have the big bucks, get paid a large fortune for it.

    What a waste of time and effort, a drain on our society. Too bad it is not directed towards something actually useful for the common good.

  6. So this company convinces a fleet of coders to write algorithms that works on this encrypted data but not on any other. The algorithms are back to the company and if, IF, the code is any good the “researchers” are payed in bitcoin?

    Why the hell the researchers would think this is even a good idea? If the whole thing is anonymous, do they even realize what their algorithms are actually doing? Due to the encryption scheme, the so-called researchers might not quite realize it.

    It’s like freelancer.com taken to an extreme.

Leave a Reply to Ken NCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.