P=1

I note that George Secor is using a sum-abs error (or P=1) criterion in
his latest post on the tuning list. Those who wish to define "poptimal"
looking only at the range 2<=P<=infinity might want to reconsider (if
they haven't already).