.jpg)
This week has been eventful when it comes to financial technology glitches. First, BATS had a minor outage due to a software update. Then on Wednesday, NYSE went out. NYSE was down for more than 3 hours. It made headlines worldwide. Even President Obama was briefed on the matter.
This week, something else was in the news – a NASA spaceship is getting ready (as of Friday morning) to do a fly-by of the former planet Pluto.
After graduating from college, I’ve mostly done two things – traded stocks and built trading systems, so I don’t know what it takes to build something that flies to the outer edges of our galaxy, but I presume it’s all quite complicated. I do know one thing – NASA doesn’t create such spaceships by magic. Someone had to sit down and write the code, create the electronics and think about everything that could possibly go wrong.
Something I find completely puzzling is how so many analysts, commentators, etc. can read about (1) humans getting photos on Facebook from Pluto, and (2) an exchange going down because they changed some of their code, and not have an issue reconciling the two. Why has no one asked themselves, “How is this possible?” Is it because order types are more complicated than spaceships? Perhaps the complex US market system (i.e. Reg NMS) with all of its subtleties makes a mission to Pluto look like freshman calculus? If you work in finance and you’d like to answer ‘Yes’, then you’re flattering yourself.
The big point everyone has been missing is that there’s a significant gap between how exchanges, dark pools, trading systems, etc. are designed/implemented and how spaceships/microprocessors are created. Even though people are now using FPGA’s, microwave towers, processing social feeds to capture ‘alpha’, and so forth, the financial services industry hasn’t changed the way people write code since computers became popular in finance a couple of decades ago. In fact, finance as an industry is quite behind the times when it comes to the science of creating reliable software systems. Complexity is there – a lot of automated decision-making everywhere you look, but the tools for managing the complexity are decades out of date. It’s not about a lack of talented developers; finance is full of incredibly smart people. It’s about the tools they have at their disposal for managing the incredible complexity of financial algorithms.
Click here to read the full blog post.
