650,000

First things first. I'm not a statistician. Even the great Worstall got the first Lancet Report spectacularly wrong when he tried to dissect it fully, and I get the feeling that on matters numerate he's a touch smarter a cookie then I am. However, something really seems wrong with that figure. 650,000 - the same number of deaths as the British army suffered in the Great War - a war where 50,000 died in a single day. 50,000 more dead than in the US Civil War. About 150,000 more than British military and civilian casualties in the Second World War. Approaching 8% of the entire population of Iraq. Ten times as many as any other estimation.

It seems to me to be a classic example of analysis, from two Greek words, 'ysis' meaning 'to pull numbers out of'.