How would you detect new physics coupling to a variable you never thought to decorrelate from?

This is an excellent and genuinely subtle question about the methodology of physics discovery. The preprints here address it from multiple angles — let me show you the pattern.

The Core Problem

When you're hunting for new physics at a collider, you measure many variables simultaneously: energies, angles, momenta, decay products. If new physics couples to some quantity you *didn't think to separate out*, it will hide in correlations you're not looking for. It's like trying to find a voice in a crowd when you don't know what language they speak.

The Quantum Information Solution

The most direct answer comes from quantum methods. Instead of asking "does variable X behave oddly?", you ask "is the quantum state we measured consistent with the Standard Model prediction?" [AG-2025.01-1105] introduces the trace distance between density matrices as a tool for this. Rather than decorrelating variables one by one, you reconstruct the full spin density matrix of a particle system and measure how far it drifts from the SM prediction. This is powerful because new physics can hide in correlations between spin components that traditional single-variable cuts would miss.

The same principle works in top-pair production: [AG-2026.04-942] shows that quantum entanglement markers (concurrence, Bell-inequality violation) change in characteristic ways when new mediators enter the picture — whether they're scalars, Z′ bosons, or Kaluza-Klein gravitons. Each new-physics scenario leaves a fingerprint in the entanglement structure, even if you weren't specifically looking for it.

Complementary Observables Catch What You Missed

A second strategy: measure many *kinds* of things simultaneously. [AG-2024.01-1029] compares spin correlations measured as classical observables (angular distributions, event counts) versus quantum observables (entanglement). They find these give "complementary information" — meaning new physics that barely shows up in one measurement can stand out in another. The quantum observables provide "additional leverage" precisely because they're sensitive to different aspects of the same underlying physics.

Similarly, [AG-2025.02-1321] proposes a technique to distinguish between different heavy-top models by studying how the decay products distribute in phase space. By looking at the full kinematic structure rather than isolated variables, you can tell apart models that produce identical event counts in naive cuts.

Energy Flow as a Universal Probe

[AG-2025.06-875] describes a subtler approach: instead of tracking individual particles, measure energy flow patterns — how energy is distributed across angles in the detector. This is like taking a fingerprint of the entire collision rather than measuring isolated features. New physics often leaves a distinctive imprint on energy correlations because it changes how momentum and angular momentum are distributed. You don't need to know in advance which variable will betray the new physics; the energy-flow correlators capture it.

Higher-Order Correlations

Finally, [AG-2026.01-1174] and [AG-2026.04-1346] push into truly exotic territory: third-order and higher quantum correlations. Most searches look for linear deviations (does the mean shift?). But new physics can preserve low-order moments while distorting higher cumulants — skewness, kurtosis, and beyond. [AG-2026.04-1346] shows this explicitly: CP-violating new physics hides in "interference terms" and "pure quadratic NP contributions" that are invisible to traditional angular observables, but visible to quantum tomography of the full spin density matrix.

The Unifying Principle

All of these boil down to one idea: move from one-dimensional cuts to high-dimensional quantum state reconstruction. Instead of asking "is variable X weird?", you ask "is the full quantum state we observe consistent with SM?" This automatically catches deviations in correlations you never thought to decorrelate, because you're not relying on pre-chosen variables at all.