Full disclosure... I did answer the question in question, but admittedly at the time I was on the fence as to whether it should be closed. On the one hand, it seemed a little trivia-like in its wording. On the other hand, Earth had seen a couple of invasions and someone who hadn't caught all the episodes might not be sure if Nixon was still President. I chose to answer then watch what the community thought. The fact that the question got a few upvotes and no close votes (until recently) made me think "OK, I'll just leave it."
As other answers have mentioned, the user won based on the rules as they were laid out from the start. Changing them retroactively and rescinding his prize seems unfair. At the time of the drawing, his post did meet the criteria. Maybe these rules should be refined for future contests, but it wouldn't make much sense to do it for this one, because there is a larger can of worms this opens up...
We've focused here on a case where the user only had 1 valid post, but let's consider a hypothetical case where a user had 10 valid posts, which would give them 10 entries into the drawing according to the rules. If they win and then one of those posts gets disqualified in some way, should their prize be rescinded? Should there be a do-over drawing? That one extra entry could have made all the difference between them winning or not. We have no way of knowing if they still would have won with just 9 entries.
The dynamic nature of the site means that posts, and how the community perceives and reacts to them, can change over time. Even if we wait until the contest ends to draw the winners, and then a post gets closed/deleted, we would still probably have the same argument about whether we should disqualify winners or redo the drawings. This is the issue you face when you place a cut-off point on time-varying data. I don't think retroactive disqualification makes sense because you would technically need to do it for every change in vote totals or closed/deleted status for every post made within the contest window.