Ciper - actually, what you said in your 3rd paragraph is precisely what happens...the ecu sees the higher octane, likes it for a time, then goes right back to where it was before. Datalogs of mine and many a customer car have verified this. As for running maximum advance, again, I can confirm via datalogs to my car, that max advance did not change., only the advance at certain rpms did. However, once the ecu figured it all out 9which does not take long mind you..in my case, around 75 miles of my normal driving, which is a mix of highway and local), then it scaled timing back to the old map.
As for the detergent issue Speedwagon, I hear what your saying, but then again, my GVR4 will literally run on whatever gas I give it (I always use premium, as its a turbo car). it does not care if its 93 or 94, and it does not care what brand...again, I have countless datalogs to show it. The only gas that madea difference was race gas (which I can run since the car is catless)...it then runs quite rich once its cycled through the lines (whcih happens after around 10 minutes of leaving the car to idle). The resultant changes in timing, etc then are the result of my having to lean the car out on the VPC at idle, and the AFC at other rpm points, whcih in turn causes a change in timing.
At the end of the day, I really just wish Subaru would hand off all their ecu work to Mitsubishi to handle......the Subaru ecu's are so finicky its sickening, and they rarely run consistenly day to day. Even the new STi is plagued by this as well (at least on some cars)
I do agree Speedwagon, that it comes down to testing what your car likes...and going with that. My disagreement was that if the ecu calls for 87, it will not run better on 93 for the reasons described above....at a certain point, the ecu scales things back. But maybe what we are finding is that the 87 is of such marginal quality, that the step up to 89 keeps things happier....could be.