Im curious. When tuning VVE, we use LTFT or lambda error to fine tune the table. Why doesn?t the ECU do this automatically? Or does it? Sorry for asking in a round about way, but does the ECM have only one pair of memory locations (bank 1 and bank 2) for learning LTFT offset? It would seem much better to have a table with these learned values averaged long term, for the ECM to use as a starting integral point which would then use live STFT to wind up or down, to effectively self tune and compensate for any VE table error. Its not likely that the LTFT offset at one operating point on the VE table is right for another point...so having the LTFT relearn all points all the time seems inefficient to me...