It's a nice simple question. And a good explanation.
However, the reference in the explanation to Rob Conery's adaption of Itzak's description is a bit unfortunate, because he recommends "a simpler mathematical workaround that avoids division altogether" which, since the columns concerned are stated to be integers so that division can lead to rounding, will deliver incorrect results as often as correct ones. For example, with default rounding, if col2 is 7 and col1 is 3, col2>2*col1 is TRUE but col2\col1>2 is FALSE, so the "simpler mathematical workaround" returns rows that the method using CASE does not return.
Also, the method Rob gives for using CASE is far more verbose and complicated than it need be, and that unneccessary complication is the only justification for presenting the non-working "workaround". (I wonder if we need a new word "doesn'tworkaround" for such things?) Here is the simple version:-
SELECT col1, col2
WHERE CASE WHEN COL1 = 0 THEN NULL ELSE Col2\Col1 END > 2;
All this doesn't really detract from the value of the question, but it certainly does detract from the value of the page referenced.