You're right that the article states, that the AES SOR would sometimes be (mis?) classified as oppotunistic - so the metric is not perfect. I'm actually incline

You're right that the article states, that the AES SOR would sometimes be (mis?) classified as oppotunistic - so the metric is not perfect. I'm actually inclined to believe that order send/cancellation rate (see analysis in eg. [1]) would be better. But the entire paragraph (in my reading, at least) seems to be arguing, that you will loose money, if you have short-term negative alpha (footnote 8 actually retracts the argument again). This is obviously not true, as this would only leave room for the HFTs in the market. [1] http://www.nanex.net/aqck2/4661.html http://www.nanex.net/aqck2/4661.html