I have been unable to find consistent OOS results using cointegration. Those pairs that did hold up wound up being super lightly traded, and when backtesting with realistic quote/spread data, the mean-reversion disappeared.
That said, when I was still looking into this stuff (and I might one day go back to it), I found the correlation test totally useless, since some of the most promising pairs were those that had lagged correlations or none at all. Aaron Brown's advice on this topic was excellent: https://www.quantopian.com/posts/how-to-build-a-pairs-trading-strategy-on-quantopian
To quote:
Look at the pairs, concentrating on the ones that seem somewhat related but not completely obvious. Don't just ask why they appear cointegrated, also ask why they deviate in the short term. Ultimately you need both forces to be strong for a robust pairs trade. Also, don't just look at times the relation worked (deviation/correction) but at times when it didn't. Most of the time you'll conclude that either the apparent cointegration or apparent deviations were random noise, discrete events not likely to be repeated or unexplainable.
