ABSTRACT
Analogy with Regression. The structural causal equation y = b0vbix [F] and its probability counterpart E[y|x] = povpix (assuming IFR) are obviously intended to mimic the phenomenally useful linear regression model y = Oo + (XiX + e and its probability form E[y|x] = OQ + otjx. Indeed, the only difference between the two is the use of unitary vs. usual addition. I hope by now the reader will appreciate some of the virtues of using unitary algebra for binary causal equations.
