This research problem falls in the broader category of pricing and market making power perps better.
- Better model implied volatility on shorter timeframes (1min -1day timeframe) based on realized volatility.
- Take into account upside/downside volatility asymmetry.
- Factor in volatility skew into power perps pricing.
- Data collection
- Historical Backtests for model selection
- Implementing model in production for entropy market-making bot
@wvu18 brought up that he has expertise in this area, and would be interested in leading this research. Will let him scope the problem out more.