This paper examines the impact of model resolution on the skill of precipitation forecasts, particularly in the context of convective events. The authors use a verification method that compares rainfall forecasts with radar data using fractional coverage over different spatial scales. The Met Office Unified Model was run at grid spacings of 12 km, 4 km, and 1 km for 10 days during the summers of 2003 and 2004, with all forecasts initialized from a 12 km resolution. The results show that the 1 km model was the most skillful over scales larger than approximately 40-70 km, while the 4 km model did not show significant improvement over the 12 km model due to difficulties in representing convection at this resolution. The greatest improvement in skill was observed for heavier, more localized rain. The paper also introduces a measure of acceptable skill, FSS_uniform, which defines the smallest scale over which the model can achieve a target skill level. The findings highlight the importance of considering both spatial scale and threshold when evaluating forecast skill and suggest that there is an optimal range of scales for model output to be useful.This paper examines the impact of model resolution on the skill of precipitation forecasts, particularly in the context of convective events. The authors use a verification method that compares rainfall forecasts with radar data using fractional coverage over different spatial scales. The Met Office Unified Model was run at grid spacings of 12 km, 4 km, and 1 km for 10 days during the summers of 2003 and 2004, with all forecasts initialized from a 12 km resolution. The results show that the 1 km model was the most skillful over scales larger than approximately 40-70 km, while the 4 km model did not show significant improvement over the 12 km model due to difficulties in representing convection at this resolution. The greatest improvement in skill was observed for heavier, more localized rain. The paper also introduces a measure of acceptable skill, FSS_uniform, which defines the smallest scale over which the model can achieve a target skill level. The findings highlight the importance of considering both spatial scale and threshold when evaluating forecast skill and suggest that there is an optimal range of scales for model output to be useful.