i simulated this strategy on excel using realtime data but couldn't get it to turn a profit. could you name the paper? i don't know what i'm missing...

you're right, errrb, longer history did make it very unstable. thx! for confirming my doubts.then my next question is, what can one do to extract a cointegrated port given m-days of history on n-stocks (m by n matrix), with short sales allowed? or is this problem too open-ended?

<t>hi. could someone please tell me what i'm doing wrong here:starting with log daily closing prices (say 150 days) of 100 stocks, i regressed on a sine wave with the constraint that the wgts have to equal 1. to my surprise, i was able to get a set of wgts for a portfolio whose nav very closely mimi...

<t>hi. could someone please tell me what i'm doing wrong here:starting with log daily closing prices (say 150 days) of 100 stocks, i regressed on a sine wave with the constraint that the wgts have to equal 1. to my surprise, i was able to get a set of wgts for the portfolio whose nav very closely mi...