We never will see such changes for anything as far out as the hypothetical planet nine. The orbital periods for objects beyond 100AU are measured in thousands of years, as far as our observations go they might as well be stationary.

Yes, unusual values for orbital parameters are only evidence that something happened in the past, but that thing that happened in the past might have been "close approach with a planet". And as a hypothesis, this is in no way less valid or likely than a flyby with a star.

We observe a clustering of orbits, they are correlated. If the object went away the clustering would degrade until there was no longer a signal.

Orbits do not change by themselves, but only under the influence of another big celestial body, when kinetic energy, momentum and angular momentum are transferred between the interacting bodies.

There is no such thing as a degradation of an orbit.

The fact that the clustering has not degraded is actually evidence for the opposite fact, that the body that has perturbed all those orbits is no longer there.

What? No. Maybe in solar systems without a Jupiter.

Apsidal precession makes the aphelion of eccentric orbits rotate around the Sun over time. The rate at which this happens depends highly on their orbital characteristics, meaning that if something perturbed their orbits to point in a specific direction, over a relatively short timescale (single digit millions of years), you would expect their aphelions to point in essentially random directions.

The fact that TNO orbits seem to cluster into specific directions is strong evidence that something is actively maintaining their orbits, by repeatedly perturbing them.

A large planet that's quite far out is a reasonable hypothesis for what's doing this. (It's not the only one.)

But were a search to _find_ planet 9, would you still be holding out for changes?

If you made a direct detection, the orbits vs changes question would become immaterial.