I've found quite a few things on these flights that made it past the ground checks.
You can't simulate air loads, or the changing geometry of the airframe, or g's.
I guess the worst I found was the CF-6 that wouldn't start inflight. Had lost a few compressor rows - but not the ones normally borescoped. The only clue - it would start on the ground but not in flight.
I've found a dump valve that stuck open. Imagine that - lose an engine, dump to landing weight (MD11), then the fuel automatics start shifting fuel to correct the imbalance and you start dumping again....
We had an APU that would not start inflight, but was fine on the ground. After SIX reflies they found an arc in the ignition circuit - the flexing of the airframe in flight caused it.
I've found two ADG's (RAT's) that did not work. They had both been spin-tested on the ground.
I guess the best point I can make is this: my company would absolutely not do these test flights if they didn't have to. They save us money, in other words.