Identifying the links between fire danger metrics and fire activity is critical in various operational and research fields. A common methodology consists in analysing the relationship between cumulative burnt areas and fire danger metrics. Building on this approach, it has been proposed that fuel moisture content (FMC) drives fire activity in some ecosystems through between one and three breakpoints corresponding to the onset or saturation of fire activity. We demonstrate, through two different approaches, that this methodology is incorrect, because it is biased by the frequency distribution of FMC values. From comparison with a neutral fire distribution and correction for the frequency bias, we show that cumulative burnt area breakpoints are spurious: an upper breakpoint might exist (but would be higher than expected), while no evidence of reduced fire danger was detected for the lowest values of FMC (on the contrary, a secondary increase was detected). Our findings clearly suggest that previous breakpoints resulting from this methodology should be considered with caution, as erroneous conclusions regarding fire danger breakpoints could have major consequences for both fire safety and science outcomes. Finally, we discuss widespread confusion between fire danger breakpoints and fire danger levels, which explains most previous erroneous conclusions.