risk_generator's profile picture. insecure underachiever. Short-term strats & HFT, NN RL ML (non critical path). Rust, python.

!(NoRisk)

@risk_generator

insecure underachiever. Short-term strats & HFT, NN RL ML (non critical path). Rust, python.

I remember having this as alarm clock


!(NoRisk) reposted

Evaluating a model on training data is like asking your mom if you look good.


Market is almost efficient. This is good. It means that the drivers are almost priced in, thus, the market data *just* needs to be untangled and analyzed in order to back out some of the (abstract) drivers. I don't even need to know what are the drivers, as long as they predict.

I've met a lot of people outside of finance who actually believe in EMH. If all information gets priced in, who prices it in? What's going on in the time between the information coming out and this person pricing the information in?



visualize 3D order book for fun: 3dorderbook.com

risk_generator's tweet image. visualize 3D order book for fun:
3dorderbook.com

02:03 Duality’s strategy and why it uses ML 11:02 How Duality uses machine learning 15:33 Interpretability and overfitting 29:40 Spot flawed ML strategies 36:46 Mean field games 52:05 Lessons from Jim Simons 54:22 Physics and finance risk.net/cutting-edge/7… #quant #ml


Loading...

Something went wrong.


Something went wrong.