Accurate flood estimation is fundamental to effective flood risk management. The most recent updates of the Australian Rainfall and Runoff: A Guide to Flood Estimation, first published as ARR2016 and refined in ARR2019, represented a significant shift in methodology in both flow-based and rainfall-based approaches. Moving away from the traditional design storm approach, these guidelines introduced a Monte Carlo framework, integrating rainfall intensities, temporal patterns, storm durations, and losses to generate frequency distributions of flows. While these developments marked a considerable step forward, subsequent research, particularly in South Australia, has exposed several challenges in applying the new framework. This paper presents key insights from UniSA’s ongoing research into the core methodologies of flood estimation. Central to this investigation is the recognition that most streamflow estimates rely on rating curves that often extrapolate well beyond observed data, introducing significant uncertainty into flood quantification. Furthermore, the performance of industry-standard runoff routing models (e.g., RORB, URBS, and WBNM) in performing rainfall-based flood estimations has been shown to vary based on several aspects, including model structure and temporal resolution, with time step selection directly affecting both storage parameters and loss estimates. Another important finding is the inadequacy of assuming constant hydrological losses, as was done in the development of ARR2019. In South Australia, distinct seasonal rainfall patterns suggest that separate loss regimes may be needed for summer and winter events. The research, supported by the Disaster Ready Fund (DRF), highlights the importance of revisiting foundational assumptions in flood estimation to enhance the accuracy of design flood estimates, thereby mitigating associated risks and fostering greater resilience in flood-prone communities.