Quants - Build Technologies That Fix What They Broke

This post is inspired by this isn't going to work in Seth Godin's Blog.

It is about taming a system that runs out of its propose. In many cases regulations or motivations will not work - but the technology it is built of may.


Take the car itself. There are a lot of rules how to behave in a traffic, situation dependent speed limits, ... and motivations like "keep right" (or left), "slow down" when it is foggy ...

But recall the time where cars became powerful but so weak in their driving behavior that driving became dangerous. About 20 years ago, I drove at a high way in a completely straight flat section. Heavy rain. Approx. 90 km / h speed. Suddenly a car  broke loose ending in a pirouette. I could stop, but with more luck than virtuosity.

Most cars have now modernized built-in technologies that support driving through curves, break correctly in all kind of conditions ... a luxury car has now up to 10 million lines of code.

In From Financial Daredevils to Autopilots I have taken the metaphor of the golden age of aviations to outline that such intrinsic technology developments often occur in jumps. Based on kind of self-organizing technologies (parts of UnRisk are developed in UnRisk).

Motivated by my assessment that great motivations, like The Financial Modeler's Manifesto or regulations can fix what the technology broke: in short, from too complicated models to uninformative when not wrong VaR treatment ...

This was 2011. And enthusiastic about our decision to unleash the programming power behind UnRisk - UnRisk-Q - I thought we will implement more of this auto-control systems. But from the feed-back cycles with featured customers and their quants, we know now a better way: enable the quants, who have their individual knowledge framework, to do all kind of auto-valuation-and-risk-control and let us concentrate on the support of these high-level task by, say, cross-model checking, scenario-building and testing, ... capabilities in UnRisk-Q.

At the moment, we are working hard on CVA/FVA/DVA, in order to provide the same flexibility as the UnRisk engines, VaR Universe, .... No we will not make this hastily.

Because, we think, "you must adjust this and value that" is not enough. It is quantitative technology that can improve quantitative technology - done by quants.