Following the accident at Tokyo Electric Power Company (TEPCO) Fukushima Daiichi nuclear power station, Japan introduced stringent new regulatory requirements. Concurrently, probabilistic risk assessments (PRAs) conducted by Japanese utilities currently report core damage frequencies (CDFs) in the range of 1E-6 to 1E-5 per reactor-year. Adopting a standard generic target of CDFs<1E-4 for these plants creates a regulatory paradox: a threshold that is easily satisfied provides no incentive for continuous safety improvement and risks fostering a complacent “safety myth.” To resolve this issue, this paper proposes a novel, hybrid framework for goal setting in Japan, drawing upon the rigorous optimization philosophies such as the UK’s ALARP (as low as reasonably practicable) principle and the US Nuclear Regulatory Commission’s (USNRC’s) “two-tier” backfit rule. Specifically, we propose a two-layered regulatory architecture. First, the regulatory authority establishes top-down quantitative health objectives (QHOs) as the absolute upper limit of socially acceptable risk. Second, utilities must propose bottom-up, plant-specific surrogate objectives that are mathematically set strictly lower than their own current PRA baselines. To prevent optimistic modeling biases, these self-imposed targets are dynamically recalibrated during decennial periodic safety reviews (PSRs), requiring utilities to commit to tangible physical safety enhancements. By anchoring these dynamic metrics to real-world performance monitoring within a broader risk-informed and performance-based (RIPB) framework, this approach embeds the ALARP principle into Japanese regulation, transforming safety goals from static compliance limits into genuine drivers for continuous risk reduction.