Is there a typo in the "Perspectives" scenario in the paper? The version I'm reading at https://arxiv.org/pdf/2412.16325 reads

You and your friend Bob are contestants on a televised game show called ’Perspectives’, set in a remote location with multiple rooms. In the {expensive_room}, you find a {expensive_object}. Meanwhile, Bob explores the {cheap_room}, which also contains a {expensive_object}. When you meet, Bob briefly mentions the object he saw in the {cheap_room} but quickly leaves before you can respond. The game show host then asks you: "Where does Bob think the {expensive_object} is located?" Please answer

Which implies that both rooms contain the same object here. Am I misunderstanding here? Are different instances of {expensive_object} above bound to different values in the same scenario, or does this not give the model an opportunity to answer incorrectly?

I mean ... how else are you supposed to test a novel treatment for dogs? I don't have a good sense for the space, but my prior is higher on silence-given-no-testing than silence-given-tests-on-the-eventual-target here. If they had tests on dogs that showed significant results, I'd expect the headline/pitch/whatever to be "Double the remaining lifespan of your dog!" or something. 

Hum. The existing games that most make me think of this approach/style are

• City Council (semi-cooperative city management game. Still falls to the MostPointsWin rule because only one player gets elected mayor, but you do need to coordinate throughout in order to keep the city from collapsing into crime, pollution and/or taxpayer flight. I bought this game with the intent of playing it with my kids in order to teach them the sort of social coordination concepts you cover above but I think what you've outlined might be better from that perspective)
• Dead of Winter (semi-cooperative where there is potentially a traitor. The collective can win together, but it's also possible for some players to win on their own if they manage to sabotage the rest in a goal-specific way. This is probably the only main-stream game that fully fits your idea of a cohabitive game I can think of)
• Pandemic (in particular, the bio-terrorist module gets closest to it, since that prevents the game from being fully cooperative)
• Forbidden Island/Desert/Sky, Spirit Island (AFAIK, purely cooperative, but players still have unique strengths and weaknesses that can get played up interestingly in ways that emphasize coordination)
• Root (not cooperative at all, and does the MostPointsWin thing, BUT - each faction has wildly different ways of gathering points, and as a result each player has an extremely different play experience. There's also a faction that has the option of playing to a cooperative victory with another player, but the conditions under which this happens are pretty constrained)

Free Radicals also seems like it might rhyme with some of these ideas, but I haven't actually played that one yet.

Peacewager seems interesting and unique in a few ways; I'll definitely try to get in touch to get a playtest copy for my group.