The counterfactual analysis aspect
A case worth reading, because it accurately explains how the but-for counterfactual hypothetical (i.e., possible, what-if, etc., world) analysis is to be done. (For those who have been paying attention, that’s the analysis for a test that we’ve been told isn’t metaphysical at all, which is one reason why we’re to use it rather than go adventuring with philosophers into abstract metaphysics, or something like that.)
Lee v Minister for Correctional Services,  ZACC 30, 2013 (1) SACR 213 (CC) (11 December 2012) starting at para. 37.
As a colleague put it, the court’s explanation is “far better than anything our Supreme Court has ever come up with”.
One case that isn’t helpful on the “how” of the counterfactual analysis – even if, for other reasons, the result of the case is correct: I make no comment on that point – is Eli Lilly and Company v. Apotex Inc., 2014 FC 1254 at paras. 20-35. I believe that, after you read Lee, if you read Lee, you’ll understand why the discussion in Eli Lilly isn’t helpful. And why the explanation is wrong, even if the conclusion is correct.
The how much evidence is enough aspect – i.e., the Snell robust and pragmatic inference
Another case worth reading is St-Germain c. Benhaim, 2014 QCCA 2207 (CanLII) – see the majority reasons of KASIRER and BÉLANGER, JJ.A (in English).