Here is the 2nd development update on Sep 2020.
During this sprint, we have finished a prediction algorithm for a big crash. The algorithm is a good achievement in our milestone because we can protect assets under management before a big financial crisis. To get details, please check the following.
Athena
Market Sensing Algorithm
- Several tests taught us that predicting a big crisis is more comfortable than a bubble. A significant financial crisis leaves a footprint in various ways before it happens. So the algorithm is designed to track related data to monitor abnormal events
- The hardest technical issues which make us miserable are placed in two areas, 1) defining abnormal event, 2) control sensitivity. Those two issues are closely correlated; it is like the chicken and egg problem.
- We applied a statistical approach(distribution) to define anomaly, and the algorithm has self-tuning function to control sensitivity.
- When we test the algorithm with actual data, it successfully detects well known financial crisis such as 2020 covid and 2008 one.
- The algorithm covers Korea stock market as of now; we will expand its coverage to US, China as well as crypto market.
- We are going to use this algorithm for risk management.
Mercury
Dashboard
- We have implemented a chart showing a risk signal with index data. Now it is easy to check market risk by just clicking a menu on the dashboard. Later we will add a notification function through slack or telegram.
- We have developed a chart library using D3 for fast execution and flexibility. We used a well-known chart library, but sooner or later, we realized that our needed functions are not supported, and it is tough to implement on top of the library.
- The below screenshot shows the risk detection chart on the dashboard.
- The red dot means critical, and the blue one is the warning signal.
Get early access to the latest updates on the MoA
Get early access to the latest updates on the MoA Investment Solutions
