I assume ehrenfest was asking for an example that satisfied the hypotheses in the quoted passage -- specifically, the only outcomes are "win" and "lose".

Alas, the page doesn't give a precise definition of "game" and "winning strategy"; without that, I couldn't really speculate. But since the article suggests the axiom of choice is needed, such games probably aren't explicitly constructible.

What kind of a game is not explicitly constructible?

Does that mean that if I get asked a question about a specific game on a test, I can assume that one player has a winning strategy? Can one prove that for explicitly constructable games?