Azimo,
Hopefully the OP has been answered for you. There are, however, a couple of points that I would like to make: -
"The cars worked well in most conditions other than full load and engine start up. During these two conditions the ECU switched to open loop causing the ECU to look into the mapping table rather than to receiver feedback from the oxygen sensor"
In a gasoline engine fuelling during start is always open loop because the sensor will not have lit off, even with a heated sensor the dewpoint of the exhaust gas must first be reached before full heating can be used.
In a gasoline engine full load, with a HEGO sensor, is also open loop - since a binary sensor will only read rich/lean and full load is 99% of the time within the enrichment area the control can only be open loop.
"Normally, ignition timing, load and rpm goes hand in hand in the mapping table. The knock sensor on the other hand will push it further to the brink of mild knocking to really push the ignition timing to the limit."
This is not how a gasoline engines ignition system works. The spark will be set at a calibrated base timing which either corresponds to MBT or x degrees from DBL (emissions/driveability permiting).
If knock is then encountered the spark is retarded from this base timing and then readvanced stepwise until the base timing is again reached. At no point will the spark advance further than that which is already precalibrated. Hence the extra RON of the methanol will not result in more spark advance.
Since the greater amount of fuel required for stoich and the lack of the control systems ability to advance the ignition past the corresponding calibrated base timing for gasoline the BSFC will suffer tremendously.
MS