@Wordle is the only right answer, though Minesweeper could be equally hilarious

Nethack would be great, might take a while but you could also do one of the variants with a Convict role

Suggestion of Battle for Wesnoth:

It's a turn based strategy game. There's a multiplayer mode, several campaigns of various difficulties, and various quite hard co-op scenarios. It has a very high branching factor due to RNG-heavy mechanics, making gameplay quite unpredictable, which is a core part of the difficulty. Hence why I think it'd be very interesting to see prediction markets play this game.

Reasoning behind my suggestions:
Slay the Spire - Meets every requirement in the description perfectly, and one of the devs is an active manifold user which is probably worth something. There are a lot of decisions in the fights which might mean the best format is only betting on card selection/world map stuff while an experienced human player does the combat (imo this is ideal), or maybe there's some way to lower the decision space for a betting market (like creating markets to choose between multiple viable decisions in combat, with the host choosing the viable decisions to bet on).

Wavelength - The game is about predicting very specific percentages, which suits a prediction market for very obvious reasons. Might not be fun if you don't know the people who are setting the percentages for each question, could maybe be avoided by pruning the questions beforehand, having custom questions, or just having the right people/good vibes on the stream. Not really difficulty focused like the market specifies, but the engine fits prediction markets so well I think it'd be really cool anyway. Would maybe be interesting to see if a prediction market could score better than a group of people who know eachother well.

A Fake Artist Goes to New York - The game involves all the players taking turns drawing lines to construct an image matching a word, but one player doesn't know the word and is trying to avoid getting caught. I think having the stream display the word and having the market predict the fake artist instead of the players, while otherwise leaving the rules unaltered, would be a fun format.

Pokémon is a great option, if you didn’t want to tackle the full-game RPG, the official multiplayer battle format VGC is incredibly deep and complex. It’s playable through Pokémon Showdown which is a free unofficial website. You’d just have to tee up a few opponents willing to play without enabling the timer, and have a team built in advance. Happy to help with the latter and be an opponent if timing permits.

+ Very difficult

+ Reasonably small action space and decision count to bet on

+ Engaging betting on dialouge options, character creation, etc.

- Playthroughs are very long. Maybe getting out of act 1 is sufficient.

Civilization lategame has like 50 decisions per turn, so it would be impractical to bet on them all. I think it can work as long as the player is decently skilled at the game and can make tactical decisions and delegate strategic ones to markets.

This game is great and a great fit for this. It's not too long or complex for a single game, which I think is good. It also has all the "good properties". It has a good amount of hidden information, too.

Catan would be great, you could include mana transfers in your trades and buy insurance against bad rolls.

I love catan, it would be very cool

Create a new account and try make mana!

I know a game which is turn based, easy to learn and hard to master. But i have negative balance and cannot start the game.

It is called Nomic.

The games of the genre have an initial set of rules which allows players to modify the rules during the game. Usually the games of the genre have a standart voting system: player proposes a rule/amendmend and other vote on whether they want to accept it (whether they feel the rule benefits their chances to win). And usually it is forbidden to make player-targeted rules.

With free choice markets it would be easy to propose rules, and with trading it would be easy to "vote".

The market creator could be the judge, which decides how to interpret conflicting rules in edge cases.

Good idea! What markets would you want to see in a Minecraft playthrough?

@ManifoldLive actually Minecraft is kinda fast paced, so that's why I sold my bet rather quickly. it might work but we'd need creative juices for it

@duck_master @ManifoldLive manifold api bot which every second looks at current % of "go forward", "left click", "mouse to the right by 5°" and so on, and executes the highest. Streams on twitch. To change what the character does you would buy one and sell all other options. All values would fluctuate around.

Resolves to NA after the dragon kill so the most active and useful users do not lose mana by betting against the previous state.

It allows to beat the game, but does not use the concept of "bet on market to gain mana". It just turns a webpage into controller.

@KongoLandwalker hmm, that might work. however, unless traders have very fast reaction times and/or the game is slowed dramatically to something like ~1 tick per second, I feel like it could very well turn into the player walking in one direction for a long while off a cliff or into the Far Lands, or something like that

Gamers Rise Up