The article mentions a small, concentrated debris field, which may suggest they were past "settle" and into "plummet."
Basics. The turn reduces your vertical component of lift (requiring a higher angle of attack to hold altitude) and also raises stall speed.
The portion of the report that says the wings were “nearly vertical” before the aircraft disappeared behind terrain from the camera, tells the tale and sets the stage for a small debris field. They never recovered from that bank most likely.
However they ended up in the bank, they either couldn’t get out of it (stall/incipient spin, aileron on the lower wing down exacerbated the stalled condition on that wing) or it got so far over there wasn’t time or altitude left to level it and recover.
Wing wagging, rapid roll, immediate crash is all classic “stall/spin on takeoff” accident terminology often seen.
Now the question is, was the CG so far aft that the pilot was trying in vain to push the nose down with full down elevator and it was still flying itself into a stalled condition?
That’s a VERY tail heavy CG condition to do that, and I’m not sure any of the numbers shown here in conjecture have been THAT far aft, but the airplane probably still would need significant forward elevator to keep the nose down for the climb.
The rest of it though, is a the classic power-on “departure/takeoff” stall/spin description.
If it won’t climb you have to push forward. If it’ll at least hold altitude you can’t add much of a turn, because losing even a small amount of the vertical component of lift means you have to trade altitude for airspeed. And you don’t have it.
If... overloaded and aft CG, they could have kept the nose just down enough to climb a little, missed all obstacles, and kept flying straight for some time, they might have been able to climb enough eventually for gentle turns, even if they had to lose a little altitude to do them, and worked their way back around for a landing. Or even flew a while to burn down some fuel.
But ... they didn’t or couldn’t. Not sure which yet.
It’s hard to believe an ATP would not know this or be able to attempt it. It might be easier to believe an ATP had the wrong climb airspeed in their head because they didn’t read the book or heed the markings on the ASI.
Even trying to stop that roll rate with a whole lot of top rudder may have ended better than letting it keep rolling. Even if he couldn’t get the nose down. Getting it to hit the ground at some lesser bank angle than letting it fall off to one side at a nearly 90 degree bank angle would have helped dissipate some energy hitting wing first. Still may not have been survivable but with some training in aerobatics or solid unusual attitude training, there may have been a slight chance to make the outcome slightly better.
Sigh. Ugly but not too unexpected when this new video data is combined with the conjecture thus far.