The paper argues that many erroneous conclusions derived from modelling are due to mistakes in logic rather than scientific methodology. The widely accepted models predicting the catastrophic consequences of carbon emissions, and suggesting how cuts by the developed world can prevent them, all ignore population growth and distribution, and such data are not used as independent variables in the global warming models. This casts doubt on the probability of the models, and even more on the suggested solutions, as an astonishingly high degree of accuracy in highly complex forecasts over a period of almost a century would be required, without which the extremely costly ‘solutions’ would be either unnecessary or insufficient. Over a 30-year period, forecasts of population are likely to be much more accurate than those for climate. Within such a period, population in the world’s poorest countries will almost double, leading to virtually all the disasters that are predicted to arise from global warming some decades later. Since many measures taken to avoid putative global warming are likely to exacerbate the more rapidly approaching dangers of population growth, it would appear logical to give more consideration to assisting the poorer countries rather than impoverishing the rich.