“Instead [statisticians] have concentrated their efforts in building more sophisticated models without regard to the ability of such models to more accurately predict real-life data,” Makridakis and Hibon write.
If a complicated theory doesn't work, develop a more complicated theory that doesn't work.
