Everyone seems to be having problems with the classically trained example, and I was until I began to type it out. Hopefully this can help someone else.

If MOST people who are not classically trained cannot recite the lyrics, that means less than half of the people who are not classically trained can. How is it not, therefore, more likely for her to be classically trained (or at least not part of the "not classically trained" group)? Because you cannot conflate the fact that it being unlikely for someone to be in a group for it to mean they are likely to be in a different group.

Putting numbers on it if 60% of the people not classically trained cannot recite, there are still 40% of them who can. Just because 90% of classically trained people can recite, does not mean she falls into that category.

To me, a more clear example of the flaw in the logic is:

Most vegetarians like brussels sprouts and most meat eaters do not. Would it be fair to say that Anna, who enjoys brussels sprouts, is likely to be vegetarian? No. Her enjoyment of brussels sprouts does not determine her preference for eating meat. Without further information, it is just as likely that she is a meat eater that likes brussels sprouts.

They would have been applicable for logic games, but it seems to be more for the logical reasoning section. Using what they call "Lawgic" (diagraming the parts of an argument) you can solve the harder logical reasoning questions.

Surprised this didn’t mention the most vulnerable part of the claim: that the cat “intentionally” knocked over the bin.

The way I immediately Clocked that Pat was not a member but then completely missed option D