The standard approach adopted in the IPCC to identify large future changes consists in stippling regions where the multi-model mean climate response is robust and statistically significant. This approach may lead to an underestimation of the risk of large changes in regions where future projections are not robust. Here we show that a more informative diagnostic to aid risk assessment is obtained by quantifying the mean forced signal-to-noise of the individual model responses. This enables us to make statements on regions where a large future change compared to year-to-year variability is plausible, regardless of whether the mean signal is robust across the ensemble. For mean precipitation changes, we find that the majority (58% in surface area) of the unmarked regions and part (18%) of the hatched regions from the AR5 maps hid climate change responses that are on average large compared to the year-to-year variability. Based on the newer CMIP6 ensemble, a considerable risk of large annual-mean precipitation changes, despite the lack of a robust projection, exist over 21% of surface land, particularly in Central America, Northern South America (including the Amazon), Central and West Africa (including parts of the Sahel) and the Maritime continent.