It will be inherently parallel and platform agnostic and capable of valuing portfolios and single instruments, consequently hybrid deal structures, with one approach.
This will be possible, because
- new computing muscles, heterogeneous CPU/GPU architectures, will provide enormous performance
- the application of general techniques, like American Montecarlo simulation, to solve a wider range of derivative models accurate and robust is becoming fast enough
- OpenCL empowers the development of inherently parallel algorithms that are platform-agnostic
- in-memory computing will become state of the art
- non-traditional data types and machine learning technologies helping to master complexity when doing comprehensive portfolio-across-scenario analytics
In general, articles about the exabyte revolution, big data, data science, ... seem to suggest that there are better ways to do your analytics.
But beware, as I have written here, better decision support does also mean more valuations. And as I have outlined in the previous post, new regulatory and business requirements introduce much more complexity to the valuation space - billions of in-time valuations might be necessary to adjust fair prices, model exposures, collaterals, ... and manage the risk of portfolios and finally the enterprise.
Additionally models need to be validated towards quality goals of financial instruments, like derivatives, related to their purpose (hedging, investment accumulation or de-cumulation, ..).
We can only master this complexity when we introduce blazingly fast schemes, like FourierCos methods for vanilla options and more - but in interplay with the new computing and development environment advances. And yes, it will be still less code for much more.
We need math to improve valuation and risk management.
If big data will write the book of future risk management, mathematics will provide the insight on market dynamics, price waves and what have you.
I hope the bridge is not too far away that brings mathematicians and data scientists together.