There's two things you might want to look at. One is the concept of how voting is interpreted across the site, the other on how answers serve as solutions. I think that these questions should be able to work, but the approach may need some extra thought.
On voting
We're talking about the actual vote scores, not the resultant reputation (which is irrelevant to this point). Voting activity is something that is rated on a global scale, not on a per question scale, and whether or not this works in that understanding. By consequence, and by design, certainly votes are comparable within the scale of a single question. This lets you compare what is a favorable answer to the specific question.
However, the site as a whole uses votes to identify the top performance. A +20 answer on one question should be of similar quality and utility as a +20 answer on a completely different question.
That's why it's important that the voting scale be universal across the site. The problem I foresee with this "one card per answer" setup is that the voting schema may not line up with the rest of the site. It depends a bit on how the expertise in understanding the strategy will correlate with how people get the answers and how much detail. Certainly, a greater answer will not simply list just the card, but will go into great detail on why it's a great fit. If voting focuses on that detail, then this problem is taken care of. Should this be lost sight of, though, then it will not bode well for the quality of the site.
On answers
This has been the subject of many blog posts, but to put it simply - the Q&A model thrives on the essence that an individual answer attempts to serve as a conclusion to the question. That's why acceptance exists and only points out a single answer. When the "solution" to a question is not in a single answer, but is the set of all answers, it becomes difficult for a reader to grasp what the whole picture is. Our format lends itself towards a complete solution within one package (one post), the lack of discrete post organization makes it complex to simply put together combinations of answers.
Think of it like asking the question "How do I bake a cake?", and your separate answers were "You need eggs", "You need flour", and similar. It's all part of a whole, and undesirable. The proper answer would be the whole process, not just a single step. That's why we encourage that answers are solutions - so that people get the big picture in one post, not across multiple posts.
Questions where there are multiple "items" among the answers, the ones that work are those where a single item is the answer. If there are multiple strategies to solve a certain problem, then any single one of those strategies can be its own answer because, on their own, they do solve the question at hand.
How this affects these questions depends - is the goal to accumulate strategy among the various cards presented individually? Or is it to find a singular card that works? If it is the former, then it conflicts with how the engine as a norm functions. It may be better not to restrict users to a single card if this is the case - rather, the user could focus on a strategy as a whole (using both specific card make-up as well as general card roles). This also ties in a bit with the previous point, as it will inspire more focus on the detail and the strategy, thus less likely to suffer problems in voting.
I should note that sometimes, a strategy really does boil down to just one card, that it is the unique synergy of that one specific card that works perfectly with the given plan. That's fine if your answer only needs that one card. This point is more to address the scenarios where multiple cards all work towards the same general strategic plan - dividing that into multiple answers is ultimately a disservice to the reader.
