The Robot, that smashed the Knight
On the 1st of August, 2012 all the minds at the New York Stock Exchange were again captured by the high-frequency trading. Knight Capital Group Inc. has lost $440 million due to the mistake of one of its robots. On the 5th of August Knight Capital Group announced $400 million in equity financing with Wall Street firms. And not more than 3 weeks ago Knight’s market cap was estimated at $3 billion.
This dramatic case shows that it’s worth carefully scrutinizing the various aspects of algorithmic trading as its volumes keep growing on stock exchanges.
In the process of evolution system are getting more and more complicated. The wildlife began with simple cells and came out with humans. On the stock exchange the intellectual algorithmic systems are taking the place of humans. The fact that the decision-making process is getting more complicated on financial markets isn’t such a challenge for the algorithmic systems. Highly technological, adaptive and self-study neural network algorithmic systems will be able to avoid such failures in future.
If we consider the market situation with broken Knight robot as divergence of quotations and quick come back, we won’t find here some crisis or indication of fragility of the system. The most flexible trees can bent down to the very ground and stand up straight after the storm. In Japan, which is well-known for earthquakes, they construct the houses with special mechanisms in the foundation which allow the buildings to vibrate and move safely.
The real problem comes when there is a “bad” robot that distorts the market data which is used by a “good” robot as well. In order to prevent such situations algorithmic trading must be the zone of companies which implement only really intellectual algorithms (mission impossible without innovative hard and soft ware and highly professional algorithm developers).
I guess that in the future the humans will participate on stock exchanges as observers who monitor the stability of algorithmic systems. On nuclear facilities there is always a team of observers that just switches of the reactor in case of emergency. But the human being is much slower than the high-frequency algorithms are – that’s why the problem of control and risk-management remains a challenge.
For Inventum the case of Knight Capital Group gave big stimulus for development and innovation, because it pointed out the need for new discoveries and inventions to make the system of algorithmic trading as stable as it’s possible.
(с) Inventum Group (Inventum Algorithmic Asset Management)
Join us on Facebook and LinkedIn
Follow us on Twitter
