I don't either. It's reasonable to assume that momentum carried it for a second or two. Add the possibility of the engines still providing thrust, and maybe a few more seconds before it started it's meandering descent.
But do you believe it climbed 3000 feet ?
I think it's too complex a problem to do in the head. It requires a computer doing a simulation. As I've said before, Boeing came up with a range of numbers to make a model for the noseless 747. The NTSB ran multiple simulations and came up with some solutions that fit the radar data and the location of the wreckage. The NTSB did say that thrust from the engines made little difference to the simulated path, whether the thrust was zero or one hundred percent.
But do you believe it climbed 3000 feet ?
It's a complex problem requiring a simulation using numerical methods on a computer. As far as I can tell the NTSB is the only one to run a simulation. If somebody else were to do the same, I'd certainly take a serious look at it. I'd also be interested if somebody could come up with a ballistic simulation that could match the radar data and wreckage location as well as the NTSB flight simulation.