But you're certainly right that the entire exercise is kind of silly, because the metrics we are measuring against are highly dubious in the first place. So even if there had been a signal, it would not have proven the case, only given cause for further investigation.

OTOH, that turns out not to matter, because even in the most generous possible interpretation of the data - according to the metric Vladimir proposed - there is no signal. Full stop. No deconstruction of metrics necessary.

- Jake

Friends come and go. Enemies accumulate.

That starting a war is the ultimate crime was a noble sentiment during the Nuremberg trials but it seems to have been lost since. Or have anyone been indicted for that? (Bush and Blair certainly should be, but that is another story.)

Both military and civilian losses are often disproportionally on the loosing side of a war, regardless of who started it. But to count as a war crime you have to target civilians, not merely kill them while bombing for soldiers amongst them (as the ban on "Launching of Projectiles and Explosives from Balloons" only lasted five years and to my knowledge was not renewed). Or perform any other of the acts described as war crimes in the Hague and Geneva conventions.

So no, it is not necessary to assume anything about who started a war to assume that number of dead civilians are in proportion to number of war criminals. There are of course other assumptions that are necessary.

Sweden's finest (and perhaps only) collaborative, leftist e-newspaper Synapze.se