Ask HN: What are some interesting problems you are aware of?

Last week in Ask HN, trivet proposed: "Teach me something new"[1]

In the same spirit, what are some interesting problems you're aware of, or you're working on?

[1]: https://news.ycombinator.com/item?id=31204611

The standard deviation of stock market returns ("volatility") varies over time. Sometimes a typical market move is 0.5% a day, and sometimes it's 2% or more a day. Other asset classes also have this property. How do you estimate this time-varying volatility? There are GARCH and stochastic volatility models, which are now often used on high-frequency returns, but volatility estimation will never be a solved problem. A related question is how estimates of time-varying volatility should affect your asset allocation.

Reply on news.ycombinator.com