low latency, also quants usually use Platt Scaling or Isotonic Regression to fix raw AI scores before feeding them into mathematical models(probability calibration).

low latency, also quants usually use Platt Scaling or Isotonic Regression to fix raw AI scores before feeding them into mathematical models(probability calibration). but jev uses Reinforcement Learning for Calibrated Decisions so you can just feed it straight up. also a lot of different things that are a bit hard to think of for now for me.