BEAMS
In the literature: open
Do short-time chaos indicators predict long-term orbit stability
In plain words
Quick tests of chaotic behavior over a few thousand turns could replace expensive long simulations if they reliably flag the orbits that eventually escape.
Precise statement
For realistic hadron collider maps, determine whether indicators computed over 1e3 to 1e4 turns (frequency-map diffusion, finite-time Lyapunov exponents, reversibility error, or machine-learned combinations) classify initial conditions stable for 1e6 to 1e8 turns with misclassification below 1 percent, and identify where in amplitude space they fail.
What would settle it
A benchmark on several lattices comparing indicator classifications with long-term tracking outcomes.