Full Idea
According to the Mill-Ramsey-Lewis account of the laws of nature, a generalisation is a law just in case it is a theorem of every true account of the actual world that achieves the best overall balance of simplicity and strength.
Gist of Idea
The MRL view says laws are the theorems of the simplest and strongest account of the world
Source
Gideon Rosen (The Limits of Contingency [2006], 08)
Book Reference
'Identity and Modality', ed/tr. MacBride,Fraser [OUP 2006], p.34
A Reaction
The obvious objection is that many of the theorems will be utterly trivial, and that is one thing that the laws of nature are not. Unless you are including 'metaphysical laws' about very very fundamental things, like objects, properties, relations.