When Machine Learning Meets Investing
“I can calculate the motion of heavenly bodies, but not the madness of people.”
This quote is from the great scientific mind Isaac Newton when he lost $3mil in the stock market in the 18th century. He probably wasn’t the first one and definitely not the last one to attack finance with scientific methods.
Quantitative finance first emerged in 1900 when Louis Bachelier published his Theory of Speculation, but it didn’t get much traction until the second half of 20th century when economists integrated stochastic methods and probability theory in finance. Very quickly, mathematical finance became a hot science that quantifies everything in finance which is, well, mainly just risk and return. Within the field is then divided into two worlds, P and Q, but that is a discussion for another day.
Then computers came along and they grew better exponentially. As technology improved, Wall Street got rid of ticker tapes and stock certificates. Then, they started to let machines do the trading (or at least some part of it). This is also called “systematic approach” as opposed to “discretionary”. Many hedge funds thrives on this approach because it gives a lower volatility (computers don’t have a lot of emotions).
Today, it seems that computers are taking over in everything. Despite how much it hates change, even Wall Street now admits the importance of computer science. Goldman Sachs and JP Morgan, for example, now require their new employees to know basic programming.Read full article →