No matter how much effort to fit amodel to data set, the model is the key.

About the attached *.sm

All the data sets are solved in the LS sense [Least Squares]. Since Legendre,
nothing else has been found better, excapt for the miscellaneous methods to
arrive at this end LS. Some years later, Gauss republished Legendre, thus LG.
But the key is the Chloesky matrix properties, that produces 'LS'.

Ant Colony [American Scientist]
Thurber, Hahn1, Kirby2 [NIST]

Were all muffed as they don't have a model, in these 4 data sets THE MODEL.

If that new category is accepted, expect more ++++,but days are not much elastic.

Pay great attention to Kirby2 about the applicable data reduction.

jmgiraud@bell.net