Exactly, on a bone stock car you will never have ARMD intervention on a smooth road on WOT.

It's a very sensitive function. Hell, things like a lighter flywheel can affect it and make it trigger in some cases.
The thresholds should be adjusted on modified cars, so that it stays OFF when WOT.

Where ARMD is doing something is for example when you have a stiff car with stiff mounts etc and you are say in 1st or 2nd gear at low rpm just rolling.
Now if you go on the throttle, and then back off the throttle sharply the entire car will rock back and forth. With ARMD it will catch that rocking and kill the resonance instantly. If you have softer mounts it is not as noticeable.
It is meant to counter drivetrain resonance. In higher gears it is mostly unnecessary if you have good traction. It sure as hell should never trigger there - if it does it is set up wrong.
If it affects your WOT power - it is set up wrong etc.

As I said before you can say the same thing about knock control. If it was calibrated wrong for the hardware and would be randomly pulling timing then turning it off you would also make gains. Does not mean the function is useless. But that seems to be the logic of some people here "I don't understand what it does/how it works - turn it off".

Why would you trust that sketchy evidence considring you can just log B_armd

"B_ar" in ME7/MED9, should be the same for later ECUs?
Just a nitpick.

it is not selective reading. i still need to test for myself (ski is 9hrs from my current living location). i have also given the info to others to test as well.
the person commented that it was "noticeably quicker" with that file during launch. tests were back to back, on vs off.

i also dont believe the other half is bullshit vs pure gold. my ski is not always 100% consistent on top speed, even on back to back runs like that. 1mph difference does happen often.

i understand your point.
i would love hard data. there is just no logger available.

Look in a P 996 turbo file for an exampel.