macrocephalopod

this is why I don't trust any ML

every time you fit a parameter in a backtest you are doing ML, just really slowly and inefficiently

You've clearly worked in marketing

Do you disagree with Prado that much of ML is just modern statistics? Ie, like any other statistical tool like linear regression (which I think is ML as well), it's useful if you dont use it irresponsibly, right?

I'm not sure he owns that view. That's what most people have been saying for the last 5+ years

Good pt, just mentioned him bc that's who I heard that opinion from. Still curious whether you agree w him

I strongly believe all of (a) ML is overhyped in finance (b) ML is useful in finance if you are careful about how/when you use it (c) Lopez de Prado has an ego bigger than the fucking Vatican and the world would be better off if he never wrote another paper.

haha I see, where do you get that from? Ive just found his texts on ML useful