List of the reviewed stuff so far, and what I would resolve yes (from the list at time of writing):
Father's mashed potatoes: Physical object, food
Joan of Arc: Physical object, person
Astral comment's section: Religion (I'll take arguments on this), website
Islamic Geometric patterns in the met: Concept, Physical object(s), (Feels like it could be painting or sculpture; I would be interested in arguments on either)
Mice Mechanisms Dementia: Academic paper, Field of study, Dataset
School: Concept
Alpha-School: In-person event/class

There were very few additional categories that felt like they were arguable; I feel pretty good about this list. I will re-read the winner to make sure my memory isn't too warped. I note that the definition of "concept" is a general idea or understanding, most reviews are picking apart quite specific things afaict; I really expected that to appear more.

can we get rulings on the reviews so far?

Is Joan of Arc filed under 'person' even thoughthe review describes itself as a review of the evidence for her miracles?

Are Mashed Potatoes an object? Islamic Geometric Patterns?

@Wott That's a fair request. I'll run through the list in a bit. (Yes, Joan article is about a person imo. I promised a broad interpretation.)

@ArtimisFowl (I've done the exercise in a stand-alone comment)

How come when I added a category it instantly jumped to the top?

@DavidSpies when you add an option, it initiates at 50%

@DavidSpies this is an UNLINKED binary market. Each of the options is independent. They do not sum to 100%. If you read the criteria carefully, it's actually possible for multiple to resolve YES.

@bens Lol I see you were gracious enough to correct it

If you add up all the probabilities, that's about 2.7 expected categories (which seems a bit overmuch, I would have expected it to be closer to 1)

@duck_master Is there an arbitrage opportunity where you just bet against everything?

@DavidSpies no, because new categories could always be added

@bens Doesn't that just help? It makes your existing no-bets even more likely to all be correct.

@DavidSpies well then that wouldn't be arbitraging, would it? that would just be betting NO on everything (which I've been doing)

@bens I think it's arbitraging because if one of the categories you bet against wins, you still compensate by winning mana on all the others. If none of them do, then you win even more mana

@duck_master Based on my judging from the released reviews; I think 2 is closer to true. Note that some of these categories are nested, some are very broad, and many of the reviews are combining or toeing the line between different sorts of things.

@TonyGao the rules literally say that books aren't allowed, so I think even if someone does try to smuggle in a book review per se Scott will probably catch it (and if he doesn't, his readers probably will)

@duck_master Yeah, I didn't read the question properly before adding it. But I think there's a good chance it might resolve to a nonzero percentage.

@ArtimisFowl in the case of a tie for winner (two reviews tie for 1st, say), then each of their categories will resolve YES? Or will resolve to 50% each?

In the case of 1 winner being at the border between two categories, will both also resolve YES? Or to 50% each?

@bens In case 1: I would resolve both to yes at 100% each (note that each category resolves independently. I put pretty low odds on a tie, as Scott seems to avoid that aggressively, so I don't think this changes too much.)

In case 2: both would resolve yes at 100% each. In particular, you're welcome to make categories of more or less specificity, and all that apply to the winning review will resolve as yes. (For example, a review of a painting would qualify as 'art' and 'painting'.)

@ArtimisFowl for which of the two scenarios?

@bens apologies; edited above.

Wishcasting ftw :)

If someone eg contrasts the Dune video game with the Dune board game and the Dune movie, would all resolve to 100 or would it be some other way?

@Mad You have it right - all would resolve to 100. I will attempt to judge each category separately, and using the broadest applicable definition. In particular, nested and orthogonal categories are welcome. (as an example, 'something dune related' and 'board game' would both resolve to 100 for a review of the dune board game, or even the review which contrasts it against the movie and video game.)