Well, the simplest model of the force of a spring is Hooke's law, that the force is proportional to the displacement (stretching) of the spring. Hooke's law is usually only true for small displacements. -Springs cannot be stretched to any length or compressed to any length, the begin to deform, etc. A Hooke's law spring will continue to oscillate forever (harmonic oscillator), since there's no friction. The simplest way to include friction is to create a 'damped oscillator', where you have a term that represents friction, air-resistance, or whatever losses you may have, as giving rise to a force proportional to velocity. A damped oscillator will also return to its original position.
So to describe what you're seeing, you'd need a more accurate description of how the force on the string changes with displacement, and a more accurate description of how the friction/damping acts on the spring. You could model this in many ways, perhaps doing an expansion of spring's force in higher powers of x (with corresponding spring constants), and the friction in higher powers of v (and friction terms).
If you then balance the forces, as one does in deriving the equations for the harmonic/damped oscillators, you get an ordinary differential equation. But unlike the simpler models, there likely isn't an analytical solution to it. It'd be a non-linear ODE. One of the characteristics of these is that they often have many different equilibrium points, in other words, depending on how the system starts out (and it may be extremely sensitive to this, which is part of chaos theory), it can converge to different stationary points. So it's not entirely surprising that it may return to different points if compressed and released at different distances.
That's basically the best "mathematical" explanation I can give you, short of doing actual calculations. Basically, you've called our bluff.

The reason why damped oscillators are taught in all the textbooks isn't because it's a great model of real-world springs, but because it's a model that happens to have a straightforward exact solution. But you see, as soon as you step off the path of neat idealizations and exactly-solvable models, even the simplest things can get quite complicated. Welcome to physics (and the forums)!