The octopus had, we are told, correctly predicted a string of world cup matches. And, indeed, Spain won.
Suppose for the sake of the argument we could satisfy ourselves that the claims about earlier matches are all true and there was no experimenter bias here, no Clever Hans phenomenon. Would one be justified in risking money on the cephalopod's next football/soccer prediction?
Most of the social science I ever learned (not to mention common sense) says no — if there's no underlying theory (and I for one don't believe in psychic octopuses) then it's rash to rely on a correlation this is likely spurious. The argument for “sometimes” holds in some systems there may be correlations between thing you can't see, for example A and B may be correlated because both are caused by C, and if you think that might be the case you might be wise to rely on the correlation even if you can't see the link,
I can't stretch that to the octopus. But it seems some folks made money off it. Will Paul move to Las Vegas?
Update: Octopus Paul is
hanging up his spurs retiring.