As for Conway's game of life in itself it is a bad example. It use two dimensions and not reversible. Two dimensions do not allow of wave-like processes existing. In addition, model must be reversible. Universe is reversible and because of this we can know our past.
As for cellular automata in general I do't think that it can be as good model as equation. But cellular automata can be very useful as a first step to "a new kind of science" - information science or information physics. IMHO, the challenge is to combine the ideas of the algorithm and the information with the principles of interference of alternatives.
As for 37r we can not "do properties" or do anything. We can just observe 37r evolution. It very simple. Rule description you can see in Wolfram book New kind of science.
It is strange for me that Wolfram himself did not notice 37R-universe similarity I've illustrated here. I was more observant than Wolfram .
About rows mass - no, big row have not big mass.
Moreover in a long time period rows expanded and fill all closed space region. It because row is not single object but rather it is a pair of connected objects: left and right board of row.
I believe that information conservation demand in reversible 37R must produce some variant of mass parameter for "rows".
But difficulty lies in a fact that 37R-wold is not differentiable.
Therefore the conservation laws in 37R are more likely to be expressed in the some form like tables than in the form of the equation...