Human-run hedge funds emerge as 2017 victor against machines

Discussion in 'Wall St. News' started by dealmaker, Jan 14, 2018.

  1. dealmaker

    dealmaker

    #FUNDS NEWS
    JANUARY 10, 2018 / 5:43 AM / 5 DAYS AGO
    Human-run hedge funds emerge as 2017 victor against machines
    Maiya Keidan

    LONDON, Jan 10 (Reuters) - Alibaba, Tencent and AAC Technologies helped Hong Kong-based Oceanwide Asset Management make gains of 43 percent last year in its China equities hedge fund.

    Selective stock-picking by Oceanwide’s team led to the long-short strategy becoming the top-performing fund in 2017, according to a list compiled by HSBC.

    Human-led stock-picking strategies, like Oceanwide, made up three of the top five best-performing hedge funds last year, showed the HSBC research.

    Two out of the top-five were computer-driven strategies, including Swiss asset manager GAM Systematic Cantab’s CCP Quantitative Fund, which made 31 percent last year.

    But while long-short hedge funds made up almost half of the top-20, strategies run by machine algorithms accounted for just 15 percent overall.

    “I’ve not been a big believer in machines because the world has changed since Trump got elected,” said one investor with $2 billion in hedge funds.

    “I think you need a stock-picker who can put into perspective what Trump is going to do or what is happening on the other side of the world.”

    So-called long-short funds, which bet on shares rising and falling, delivered performance of 13.2 percent in 2017, when the average hedge fund made 8.54 percent, showed data from industry tracker Hedge Fund Research.

    Two of the other other hedge funds in HSBC’s top-five were run by London-based long-short equities firm Sloane Robinson.

    Five of Sloane Robinson’s stock-picking strategies, which bet on emerging markets, Japan, frontier markets and globally, made it into HSBC’s 20 top-performing funds.

    Activist, event-driven and financials-focused strategies run by humans also made up the top-20 performers. (Reporting by Maiya Keidan Editing by Jeremy Gaunt)

    https://www.reuters.com/article/hed...L8N1P52MA?mc_cid=386030a31e&mc_eid=d03fd1041f
     
  2. How can a machine do any better than to buy near the bottom and sell near the top, which humans have been doing since the first markets opened?

    The funds need to fire all the analysts and PhDs, cancel the BB, delete the algos and just hire all the great traders here on ET!!:)
     
    Last edited: Jan 14, 2018
    yiehom, volpri, Gambit and 1 other person like this.
  3. lukepoga

    lukepoga

    It’s more likely the case of people taking credit for computer strategies.
     
  4. Computers can do almost anything that humans can do when its comes to trading and they don't get tired or lose concentration so therefore can take advantage of all opportunities that present themselves.

    It's impossible for humans to execute trades with the same level of speed and accuracy as a computer algorithm especially when there's mathematics and multiple time-series involved.
     
  5. henry76

    henry76

    I don't think speed is such a huge issue when trading using fundementals which these funds seem to be doing.
     
  6. But they can't judge how much is too much or too little, which is required so often and can differ case by case. And it's that 'ALMOST' that makes the difference and keeps humans on top!!
    Buying near the bottom does not require split second timing.
    If you want to use mathematics, fine. But it's not required.

    And you mentioned accuracy. But the machines can't even draw a simple trend line correctly more than 3 times out of 10!!

    The machines might be great for crunching numbers and doing certain types of screening and for HFT. But they still lack the judgement that's required for trading.

    And, maybe most importantly, the human trader can adapt to any situation, to where the machine is limited by its program.

    Modern inventions are great. But you can't roast a turkey in a microwave!
     
    Last edited: Jan 15, 2018
    spread'em and QuasWexExort like this.
  7. Cause machines funds managers don`t know algo trading is gambling and so the appropriate variables apply.
     
  8. Is this really surprising? The machines' advantages compared to humans are basically limited to reaction time, at least for now.
     
  9. henry76

    henry76

    although they can mine vast amounts of data for certain sorts of probabilities
     
  10. sle

    sle

    • quant funds are managing only about 10-15% of total AUM in the hedge fund space so I would be surprised if they were massively represented
    • most long-short funds are, in reality, closet longs and usually do better when the market is rallying while quant funds tend to be truly market neutral
    • overall, it’s unclear what the ranking criteria is - there are funds that have done way better than the ones in the “top 20”
    • I think it will be “a few years” before quant funds would be able to compete with human-driven when it comes to fundamental long short. Really good LS guys are very good at synthetic thinking that’s hard to represent as an algorithm
    PS. I hate autocorrect!
     
    Last edited: Jan 15, 2018
    #10     Jan 15, 2018
    spread'em likes this.