What I Missed When Reading Quantitative Portfolio Management The First Time (Taylor's Version)
+ some python code
the article is on the longer side. as usual giving access to some who RT
@macrocephalopod @PodQuant they ran another behavioural experiment back in 2016 on sizing with a (known) biased coin: https://t.co/xoK0tryulV
https://t.co/ab7IyRwjW1
@BobEUnlimited@George_Saghir and btw, if macro trading was only about interpreting (public) eco data and making linkages, economists would be richer than they are
@BobEUnlimited@George_Saghir - labour market data has disappointed relative to consensus
- Fed has said they only care about this at this juncture
- you’re assuming 2s in mid-july were ‘fair’ to begin with
its a stark move but this thinking is why (macro/price) trend strats still exist
@EntropyChase if you’re minimising SSE and your variables are uncorrelated (i.e. beta = 0), then r_pred = mean. in this context, SSE cannot be greater than TSS and hence 0 <= r^2 <= 1
@EntropyChase yes can be negative if you’re running a constrained regression (eg fixing the intercept) but for any unconstrained regression r^2 will always be >=0