I have been using the new version of PTF (3.01) for about two weeks.
The interface is essentially unchanged from prior versions (the last I used before the 2.99 demo was 2.89, and it is basically the same).
The inclusion of co-integration is nice, but I am still trying to relate the current co-integration value to the historical performance of a given pair. Not much luck so far. Since the co-integration value changes over time, the only way to really look at this analytically is to have the co-integration metric at the time each trade was signaled, and compare returns from high and low scoring pairs.
So far paper trading with the new vrsion of PTF, I like it better this time around, and it is delivering on its promise.
One thing I suggested to Jared was that in the home list of pairs, one can set up signals to be delivered only if the pair meets key criteria (e.g., greater than 2 SD from the mean, along with a co-integration of 0.95 or better). That way only "good" trades are suggested from a larger list of potentially profitable pairs.