Forecasting precipitation in the tropics remains a substancial issue. Due to the stochastic nature of convection, predictions need to be probabilistic for all lead times. The coupling of convection with planetary waves calls for high‑resolution and (near‑) global coverage at the same time. Following the results of the SFB waves to weather, we aim to create a systematic comparison of available precipitation forecasts in terms of skill and cost, as well as further developing cost-efficient models.