The same finite rule handles a signed inversion start. Exact arithmetic here means exact results for the stated model inputs; measured inputs still carry uncertainty and significant-figure limits.
highlighted = computed this step
Inversion recovery starts below zero
The toy inversion start is -16 and the equilibrium is 16, so the signed gap is 32.
16−(−16)=32
A half step lands on the training null
Half of 32 is 16, so the final index is exactly 0.