Feedback Systems are inherently complex. When you introduce human inputs specially of high amplitude the results can be dangerous.
Anyway when you don' t know for sure all the "inputs" the PF "received" it's risky to conclude.
We are talking of a complex scenario with a crew (PF/PNF) lacking the training for the situation.
I love the KISS approach but for this analysis we must consider everything plausible and PF was reacting to stabilize.
Unfortunately due CRM, "tunnel effect", difficulties due "auto THS", use of SS or just panic then the entire crew failed to timely act properly or even "understand".
I understand MB comments on oscillation as a possible complex factor. OTOH in general i agree with many points OC put.
PS
I tend to respect
too much all kind of oscillations or "transients" leading to dangerous situations or crashes, mostly the ones affected by human beings inside the feedback loop introducing extra "inputs".
Remember me the "butterfly region" . I feel "butterflies" in stomach when imagining what can happen.
In general, "laissez faire" is safer, i agree. But in some cases you must act decisively.
Which ones? How to manage in "degrading modes" not perfectly timely understood.
Acting decisively avoided me two serious car crashes. And subsequently i had to stabilize the vehicle. And was very difficult and requiring extremely dangerous maneuvers. Full of adrenaline in the end.
May be applying precise and ABSOLUTELY NECESSARY inputs, proportional to the surprise in the beginning.
When confused, better to wait and see.
This remembers me about people*:
- Intelligent and active, no problem
- Intelligent and passive, no problem.
- Dumb and passive, no problem.
- Dumb and active, extremely dangerous...
(*) A friend told me about people entering the Army.