But you can multiply mean deviation by 5⁄4 (technically sqrt(pi/2)) to get standard deviation, if the distribution is normal. I.e. the two measures are identical but mean deviation is far more intuitive and therefore likely to yield better guesstimates, no?
They’re not equivalent if the distribution is skewed, because if you use a Laplace-distributed prediction, you maximize accuracy by guessing at the median value, rather than at the mean value.
Agreed, More importantly the two distribution have different kurtosis, so their tails are very different a few sigmas away
I do think the Laplace distribution is a better beginner distribution because of its fat tails, but advocating for people to use a distribution they have never heard of seems like a to tough sell :)
Any thoughts on mean deviation? Seems more intuitive in some ways
Mean deviation corresponds to Laplace-distributed predictions, rather than normally-distributed predictions.
But you can multiply mean deviation by 5⁄4 (technically sqrt(pi/2)) to get standard deviation, if the distribution is normal. I.e. the two measures are identical but mean deviation is far more intuitive and therefore likely to yield better guesstimates, no?
They’re not equivalent if the distribution is skewed, because if you use a Laplace-distributed prediction, you maximize accuracy by guessing at the median value, rather than at the mean value.
Agreed, More importantly the two distribution have different kurtosis, so their tails are very different a few sigmas away
I do think the Laplace distribution is a better beginner distribution because of its fat tails, but advocating for people to use a distribution they have never heard of seems like a to tough sell :)