OneTermSqrt #
Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.
A one-term square-root tail conversion #
This module supplies the model-free tail conversion used by the corrected Section 3.2 base-case argument. If a nonnegative observable satisfies
almost everywhere and Y has a one-sided Gamma_1 tail, taking square roots
produces a Gamma_2 random leg. There are two useful local transformations.
The unshifted form treats sqrt D as a second Gamma_2 leg and pays the
explicit binary addition constant. The shifted form absorbs sqrt D into a deterministic
shift and keeps the random square-root scale unchanged.
The exact binary-addition multiplier at Gamma_2.
Equations
Instances For
The Gamma_2 binary-addition multiplier is positive.
A deterministic constant bounded by a nonnegative scale has a one-sided
Gamma_sigma tail at that scale.
A one-random-leg square bound gives an unshifted one-sided Gamma_2
bound. Both square-root amplitudes are bounded by B, so their binary merge
has the exact scale gammaTwoAdditionConst * B.
A one-random-leg square bound gives a deterministically shifted
one-sided Gamma_2 bound. The deterministic square root is absorbed by b,
so the random square-root scale B is retained without the binary-merge
multiplier.