Edit:  I properly rescaled KMFIOP and did KFMIRL to match it as inverse and this vastly improved drivability and even boost, however it did not fix my shift issue. Since the issue only happens in AMAX im pretty sure theres nothing I can do on the ecu side. I will keep digging

The _slope_ of the line reporting torque needs to match the _slope_ of the actual torque. That is, if torque increases by 2x from point A to point B, the reported value also must increase by 2x or DSG is sad. Which usually means reported torque needs to be a line, or pretty close to a line, unless you're running your setup in some terrible inefficient area.

If you scale the whole line (ie underreport torque by 30% ACROSS THE BOARD), things continue to work as long as you re-adapt the DSG, although there is usually no reason to do this if you are competent and/or have access to a DSG ASW torque cap patch. If you flatten the top off (like most clueless tuners), everything breaks.

The absolute numeric values are not important on DQ250, the slope/gradient is what is important. The way the clutch model on DQ250 works is basically by picking two data points and drawing a line to determine the relationship between clutch engagement and pressure. The values that are used to draw the line are not important so long as everything is scaled to fit on that same line.

surely the same rescaling of dashpot, idle PID etc is required?