Serbian indictments vs. other people's dead civilians: 100/(4500+33000+2000+500) ~ 1/400
Serbian indictments times dead Serbian civilians: 100*(2300+3600) ~ 6*10^5
Croat indictments vs. other people's dead civilians: 29/(2300+33000+3600+500) ~ 1/1200
Croat indictments times dead Croat civilians: 29*(4500+2000) ~ 15*10^5
Bosnian indictments vs. other people's dead civilians: 8/(4500+2300+2000+500) ~ 1/1000
Bosnian indictments times dead Bosnian civilians: 8*33000 ~ 2.5*10^5
I'm not sure what to do with the Albanians, because I think they're from a separate, later round of wars.
Just from looking at these figures, you can see that they aren't conclusive (higher figures means greater likelihood of bias against that faction): By the first measure, Serbians get a short shrift, while Croatians appear favoured, but by the other measure, Croats get shafted and Bosniacs appear favoured. And all of these figures are well within an order of magnitude of each other, which is almost certainly an optimistic confidence interval for such a simple back-of-the-envelope calculation.
- Jake If you only spend 20 minutes of the rest of your life on economics, go spend them here.
But then, the only thing I was trying to establish with my little back-of-the-envelope calculation is that a 1:3 ratio between Serb and Croat indictees isn't completely outrageous when you look at the casualty figures.
Here is a comparison trying to accredit civilian deads, 'generously' assuming that 20% of the ethnic Bosniak (Muslim) civilian dead were killed by ethnic Croat militias and 80% by Serb ones (I suspect the ratio may be even more tilted), splitting Bosnian Serb civilian dead between Croats and Bosniaks (Muslims), and Bosnian Croat (Migeru's "Hercegovine") dead between Bosniaks (Muslims) and Serbs.
Serbian indictments vs. other people's dead civilians: 100/(4500+26400+500+950) = 1/323.5
Croat indictments vs. other people's dead civilians: 29/(2300+6600+1800) ~ 1/335
Bosnian indictments vs. other people's dead civilians: 8/(1800+950) ~ 1/340
Surprisingly close. *Lunatic*, n. One whose delusions are out of fashion.
Your indictees vs. other people's civilians figure maintained the 1:3 ratio :-)
No, it demonstrated that the results were unstable by up to an order of magnitude (one favoured Serbians over Croats by half an order of magnitude, the other the other way round), depending on which metric one uses. Which means that, pending a more detailed analysis - which it wasn't my job to do, since I wasn't trying to prove anything - any ratio below an order of magnitude in difference is not inherently suspicious.
I used two simple metrics in order to get a ballpark figure for the sensitivity to choice between simple metrics, and demonstrate that Val's simple metric was well within the sensitivity to choice of metric.
I can't really comment on your analysis, because it uses assumptions derived from knowledge of the general sequence of events during the war, which I don't know anything about.
