Renaissance Technologies’ flagship fund has earned $100bn in trading profits
THE BEST investors’ strategies often sound simple. “Whether it’s socks or stocks, I like buying quality merchandise when it’s marked down,” says Warren Buffett. Betting big on the fallout from epoch-making events, like the fall of the Berlin Wall, is George Soros’s preferred tactic. Jim Simons, the founder of Renaissance Technologies, a hedge fund, spots patterns.
顶尖投资者的策略听起来往往很简单。沃伦·巴菲特说：“我喜欢在减价的时候买优质商品，不管是袜子还是股票。”乔治·索罗斯喜欢对柏林墙倒塌等划时代事件的影响押下重注。对冲基金文艺复兴科技公司（Renaissance Technologies）的创始人吉姆·西蒙斯（Jim Simons）则是靠发现规律。
Mr Simons is less famous than Mr Soros or Mr Buffett, but no less successful. He founded Renaissance in 1982, aged 44, after a successful career in mathematics and code-breaking. Its flagship Medallion fund has earned $100bn in trading profits since 1988, mostly for its employees. The average annual return of 66% before fees makes Mr Simons one of the most successful investors of all time. He is now worth $21bn.
A new book, “The Man Who Solved the Market” by Gregory Zuckerman of the Wall Street Journal, asks how he did it. It is a compelling read. Mr Simons started investing in 1978 by looking for patterns in currencies. He had early successes with simple “reversion to the mean” strategies, buying when a currency fell far enough below its recent average. A decade later René Carmona, another mathematician, convinced him that rather than searching for such patterns themselves, they should hand over the job to an algorithm, and trade even when the logic was unclear to its human minders. In the 1990s Robert Mercer and Peter Brown, formerly of IBM, developed a “self-correcting” version of this trading approach that would double down on successful strategies and cut losing ones. These techniques, now called machine learning, have become widespread.
《华尔街日报》记者格雷戈里·祖克曼（Gregory Zuckerman）的新著《破解市场的人》（The Man Who Solved the Market）想要探究西蒙斯成功的秘诀。本书读来引人入胜。西蒙斯于1978年开始通过探寻货币规律进行投资。他在早期通过简单的“回归均值”策略（在一种货币的汇率跌到远低于其近期平均水平时买入）获得了一些成功。十年后，另一位数学家雷内·卡莫纳（René Carmona）说服他用算法替代人去寻找规律，即便是在人脑还搞不懂其中逻辑时就做交易。上世纪90年代，曾在IBM就职的罗伯特·默瑟（Robert Mercer）和彼得·布朗（Peter Brown）开发了这种交易方法的“自校正”版本，对已经取得成功的策略加大投资，反之则减少。这些技术现在被称作机器学习，已经广泛普及。
There were missteps along the way. Early in his career Mr Simons unintentionally almost cornered the market for Maine potatoes, only realising when regulators reprimanded him. For months the team struggled to make money from trading shares, until a young programmer spotted that Mr Mercer had typed a fixed value for the S&P 500 index in one of half a million lines of code, rather than getting the program to use the index’s current value.
As Mr Zuckerman lucidly explains, such strategies have limitations. One is that their scale is limited. Medallion, which trades on short-term price signals, has never held more than $10bn. The narrower the time frame, the larger the market inefficiencies and the greater the chance that an algorithm’s choice of trade will succeed. But short-termism reduces capacity. Renaissance now has funds, open to outsiders, that trade over longer horizons. But returns have been less impressive.
Other firms now try to copy Renaissance’s trades. Insiders say it tries to trade a pattern “to capacity”, moving prices so that other firms cannot spot the same signals—rather as if a bargain-hunter, upon learning that a favourite shop was holding a sale, arrived early and bought up the entire stock so that no one else even realised the sale was on. Others on Wall Street often describe Renaissance as a money-printing machine, but Mr Zuckerman shows how it has had to keep adapting its model to stay ahead of the competition.
The book’s only disappointment is that the man at the centre of it all features relatively little. That is perhaps unsurprising. Mr Simons studiously avoids publicity. After all, keeping its funds’ strategies secret is a big part of Renaissance’s success. Having solved the market, he is hardly about to give away his edge that easily. ■