anyone else struggling with cointegration vectors that break down quickly in out-of-sample backtests? I'm currently backtesting a stat arb...
Separate names with a comma.