- Australian Speaker
Kris leads machine learning research at proprietary trading firm Sortino Partners. Trained as an engineer and scientist, he has been applying machine learning to the financial markets since before it was cool to do so. Prior to becoming a partner at his current firm, Kris worked as a hedge fund quant and consulted to both buy- and sell-side on quantitative analytics and the strategic adoption of machine learning technology. Kris is passionate about a fair go for independent, non-professional traders and investors. To level the playing field, he founded robotwealth.com, an online community empowering independent traders with the practical knowledge and tools one finds in professional trading firms.
|Posted: 31 May, 2018||Subjects: Trading _Strategies Quant_Analysis&Trading|
|Source: C2018||Available to: Conference Delegates|
Researching and developing systematic trading strategies is difficult and time-consuming. There exists a vast universe of possible approaches to the markets, and traders can spend significant amounts of time performing research that yields little of value. While the chronic rejection of one's ideas is a way of life for trading systems researchers (only a small percentage of ideas ever see production), a principled, objective approach to early idea evaluation ensures that time and effort is reserved for ideas most likely to yield performant trading strategies.
Here we present a quantitative work-flow for fast, informative and robust feedback on early-stage trading ideas. The work-flow interfaces seamlessly with several data sources (local and remote), leverages modern high-performance computing techniques, and provides objective, quantitative feedback using sound statistical analysis. Best of all, the work-flow's infrastructure is built on free and open source software, and it can be used to evaluate any trading strategy whose signals can be computed prior to its simulation, making it particularly relevant to TA-based trading strategies.