(For what it’s worth I’ve found that by doing lots of commenting experiments on LessWrong it’s actually possible to develop a fairly detailed model of LW karma attribution habits, even to the point of, say, noticing which geographic regions have which voting habits. This is sorta costly, maybe, but makes “karma is like X” approximations unnecessary.)
(For what it’s worth I’ve found that by doing lots of commenting experiments on LessWrong it’s actually possible to develop a fairly detailed model of LW karma attribution habits, even to the point of, say, noticing which geographic regions have which voting habits. This is sorta costly, maybe, but makes “karma is like X” approximations unnecessary.)