My experience has been that equity prices only show a small amount of cointegration, meaning that the pair vlaue (beta1*P1 - beta2*p1) still looks, for the most part, like a random walk. Has anyone observed or been able able exploit 'highly' co-integrated pairs?