The toluene works to reduce the knock. Some days no knock , other days there is still some but small.
So I decided to do what I though would be a simple thing - retard the ignition a little where the knock retarding occurs which is around 2100 at WOT (95% load). Then I'd be able to use the crappy aussie 98 ron, which is not always 98. I mean the car goes well, but the logs show that once knock is detected in the low revs,  the retarded timing travels through the whole rev range for a few seconds and chops up to 0.5s to 60mph.

This is just a stock r32 from 2008, so nothing much compared what you guys are trying to achieve on here.

Initially I used LAMFA to make the the fuel richer at WOT. It worked wonderfully some days, then awfully on others. When it didnt work, the ignition retard was huge, 6.8 on all cylinders. I recently found a map that shows how the ecu retards more if the knock occurs with a rich mixture, so no wonder I was getting those big numbers. So back to stock LAMFA I went and I gave retarding the ignition a go.

To cut a long story short, I found that the 032KG has 4 kfzw and 4 kfzw2 maps and some kfzwa maps.  I  cut 6 degrees from 60% load up between   1500 to 3000 rpm on all the 8 maps mentioned but left the kfzwa maps alone as they are apparently not used normally.

And when I tested in the car, the  ignition timing logged by vcds in that area  is  the same as before.This is the the 8th time I've flashed the ecu just to try changing the timing, changing each of those maps at a time and  I guess next will be those kfzwa maps, cos I cant see any others!
Is there some trick to acieving this simple task in this overly engineered ecu? Why the F is the timing not following the damn maps.

The actual load reported by vcds is in the mid to high 90s. Throttle is 99% open, the timing though is around 16 to 17  at 2100 rpm. About 6 to 7 more than the maps I modified.