trading-hackers
Tail risk - the stuff that kills portfolios. Getty

Markets are not governed by an unalterable rationality; rather they reflect the humanity of their participants. Thanks to big data, machine learning, not to mention the relative cheapness of computing power, we no longer have to fit everything into one efficient market model.

This is the view of Jeremy Sosabowski PhD, CEO & co-founder, AlgoDynamix. His area of expertise is identifying endogenous portfolio risk. "The risk we look at is internally generated risk. It's the panic. It's not the external stuff. It's not the Swiss central bank de-pegging their currency."

AI/ DON'T USE THIS ONE
Newsweek is hosting an AI and Data Science in Capital Markets conference on December 6-7 in New York. NewsweekMediaGroup

"I'm talking about tail risk - the lethal stuff that kills your portfolio, and if you are not careful it's going to kill your bank. And it's the toughest one to model by far."

AlgoDynamix provides a risk analytics engine based on deep data algorithms, scanning in real-time multiple data sources. These algorithms analyse the dynamic behaviour of market participants – buyers and sellers – to identify anomalous clusters of behaviour. It's a solution being sold to banks, hedge funds and family offices.

Sosabowski believes historical data is interesting but relying on it is the wrong way to approach financial time series. "Our platform and our insights will tell you that something is about to go really, really bad. I don't like the term predictive, this is not predictive; it's forecasting to some extent but it's already there.

"There are already problems in the time series, something is already happening, panic is building up, you just can't see it yet. You can't see it yet because you don't have the deep data insight tools. We are giving you those insights."

A good way to illustrate internally generated risk is the Millennium Bridge example. Its designers and engineers overlooked the fact that pedestrians crossing a bridge that has lateral sway have an unconscious tendency to match their footsteps to the sway, exacerbating it. This led to the bridge being closed.

"They'd designed the bridge on the assumption that people would walk randomly on the bridge," said Sosabowski. "Random markets, random data, that's what you assume; Brownian Motion, that sort of stuff.

"I'm not a big fan of Brownian Motion or statistical distributions, you probably gather. Nobody in Cambridge on our team believes that distribution or Brownian Motion or statistics is the right tool set for financial data."

Sosabowski said these events are not one-off black swans. These are the grey swans that typically occur about 20 or 25 times a year. "When these sorts of things happen, correlation tends to spike, so you think you have the diversified portfolio, you think you are covered, but when you most need it, is when things actually stop working - so that's the painful reality."

The way groups of anonymous market participants – BlackRock, Fidelity etc – are structuring their deals leaves a unique fingerprint in the order book in any normal day. The Alogodynamix solution looks for large clusters of sellers all behaving in the same way - the stampede.

In terms of a time horizon, Sosabowski said it depends on the asset class but it's typically sort of 10, 12, 14 hours ahead of time at least. "So there's a big enough window if you want to go out there and hedge. This is not particularly intra-day. This is the big moves. This is the larger money pots getting ready."

He said the platform is plugged into most of the big futures exchanges - Eurex, CME, Cebot - looking at all their futures. Regarding FX futures, he said: "I'm going to be a bit incorrect and I'm going to stereotype. North American markets tend to cluster enormously, lots of panic; same with the Nikkei Futures. The FTSE Futures, Eurex - pretty boring. There must be a sign up saying, 'don't panic'. FTSE never panics. I don't know what it is with the UK."

Interestingly, the company originally applied its mathematics to medical data, and later migrated to finance. "We started off as a medical company, looking at cardiac arrests. At some point we figured out computing power was getting so cheap we can actually put some tick data in this. That was about two and a half years ago.

"We are happy to be in banking. You might think banking is heavily regulated, try doing something in the medical space. That's regulated. Banking is fine."

This article was first published on May 13, 2016.