It seems too few people realize that “climate” is a statistical concept, and it follows that any statement about climate is a statistical assertion with all these messy things attached like probability distribution, mean and median values, confidence interval, confidence level, etc. Statistics is inherently messy stuff, but it becomes a wild jungle when applied to real-world systems where nobody can be sure that this or that sample are drawn from the same universe; there always present doubts about homogeneity of data, about statistical stability of the system which is underlying assumption of the most statistical methods, but the one which can not be verified for a system existing in only one exemplar and which can not manipulated in a laboratory. I was trained to be a rocket scientist, have master degree in fluid dynamics and physicochemical thermodynamics (official name of my specialization), and also in pure and applied mathematics. By reviewing and translating scientific papers on climate change for 5 years for a popular journal (my present job), I came to the conclusion that climatology is not a rocket science: it is much worse.