Feedback: What is an irreducibly simple model?


Feedback is our weekly column of bizarre stories, implausible advertising claims, confusing instructions and more


A COLLEAGUE forwards a paper with a most wonderful title: "Why models run hot: results from an irreducibly simple climate model" (doi.org/znq). The paper's authors are Christopher Monckton, with whom we have history; Willie Soon, also well known for his belief that climate change is nothing to do with us; David Legates; and William Briggs.


We had barely reached the end of the title when it occurred to us to take an information-theoretic approach. An "irreducibly simple model" would be one that is encoded by one bit of data. That one bit could represent your preconception of what the model's output should be – "true" or "false": take your pick.


The paper appears in Science Bulletin, whose website address is scibull.com – a choice the publishers may wish to reconsider.


Alex Delimata ...


To continue reading this article, subscribe to receive access to all of newscientist.com, including 20 years of archive content.