If I were looking for a law of life, I might look at an information theoretic approach. It would have to know how to count the bits which could be the granularity of a measurement and at least two objects, the life and its environment. The granularity would allow associating the 'causation' between bits in life and bits in the environment. Finally both life and the environment have an entropic state which is a kind of potential for the definition of the information possessed in life and the environment.

Life perhaps arises in this view of natural selection if the life can tame the environment. In other words, life can increase the entropy of the environment from its perspective faster than the environment can reduce the entropy of life from its perspective. In other words life learns how its environment is simple and the environment has a hard time being complex enough (having enough causal bits) to prevent that.

This would be a computational approach to figuring the probability of life. You need a complex enough environment and a way to count the bits of causally significant information, relatively speaking. Of course, this would mean that successful Robots would count as life.

And of course, you would need a way to gauge in deceit into the puzzles of causality, in both directions.