Locally weighted regressions (LWR) were calculated between MIL-PRF-24385 1.8 m diameter pool fire extinction results and various 19 cm pool fire metrics, as compiled in a recent database. The goal: to define correlations between bench and large-scale foam fire extinction experiments to improve bench-scale experiments to more rapidly screen the firefighting potential of environmentally-friendly foams. Bench-scale metrics included the area under the curve (AUC) for extinction time versus foam flow rate profiles, the average extinction times at binned foam flow rate ranges, and CO2-based extinction times, determined through tunable diode laser absorption spectroscopy (TDLAS) measurements above the pool fire. When cross-validation was implemented within the LWR algorithm, the AUC values, CO2-based extinction times, and extinction times for a flow rate-bin > 1500 ml/min showed the best correlations to large-scale data; however, the R2 values were all below 0.6, indicating a low level of correlation. Although differences between active versus passive firefighter and foam generation methods exist between the two scales, data accuracy and limited data are likely contributing the most to low R2 values. The LWR correlations are promising and can be improved with accurate data, providing confidence in bench-scale testing despite diminished radiation intensity and turbulence associated with larger fires.