What daily community game will we add to our website?
We are looking to experiment with adding a fun game that utilising prediction markets. What game are we most likely to add?
I like Chess since it's simple, widely known, and some Manifolk are quite good at it (I believe @mqp is rated 2100+) It could be pretty simple to do eg "If we play Qc3, will we win this game?" for a few candidate moves each round, using Futarchy to decide what to play next against an AI
I learned some lessons from https://manifold.markets/Undox/communal-chess-game-will-white-win Lessons learned: 1. It is probably good to have some code tracking the official board position, since people may have differing ideas about what that is. 2. I made it initially one move total per person, but then the game grinds to a halt, I then made it unlimited moves per person, but 2 consecutive moves not allows. This was better but it allowed two people to collude to easily. So I would say you can only make a move if the last 2 moves are not yours.
Trying to figure out the murderer is a common way to read detective novels and the like. Could work with a webcomic, fanfic, webseries, etc. Every day the author reveals a new story segment, adds a new market, and maybe resolves a market or two. Use a tag or author link to make it easy to see all markets for the story.
Here's how this could work: 1. The system generates a secret game board. 2. (for the geeks:) The system publishes `commitment = hash(nonce + gameboard)` 3. Conceptually, a game board is equivalent to 81 markets of the form "Is there a ship at coordinate A1" ... "Is there a ship at coordinate I9". In practice, the UI will probably represent it as one single game. 4. Players bet on these markets. 5. Every hour (or day, or ...) the market with the highest volume gets resolved. This is equivalent to calling a "shot" in the game. 6. Players can continue to place more bets on the other markets. For example, if part of a ship was revealed, they might bet on adjacent positions. 7. (for the geeks:) Once all markets are resolved, the system publishes the `nonce`, so that everyone can verify that it didn't cheat or favor some player. For info about the original game, refer to https://en.wikipedia.org/wiki/Battleship_(game)
Actually, a move is revealed every 15 minutes, so that the game finishes in about 20 hours. If the system then starts the next game immediately, start times will cycle nicely through all time zones, giving everyone around the globe the same conditions, and the chance to play the early game or endgame.
I think this idea is a lot better than the current options because it uses prediction markets for forecasting instead of voting. However, this would only work if the system places the ships using a non-random strategy. If the ships are randomly placed, this would just be gambling.
I could see this being a fun one-time thing but what we are really after is some sort of game that will draw users to our website daily for ~5mins (think wordle). Another concern I have over this proposition is having 81 markets created will be quite daunting for most users to interact with.
The idea here is you release a survey to registered users, and then people bet on the outcome. Could be either resolve PROB based on proportions of votes, or resolve to the more popular choice. The survey could be categorical (which outcome will happen) or linear (what probability would you assign to X) This might make long term outcome betting more fun as you can bet on what people thing the probability is, and get a sooner resolution (although wrong-er... admittedly).
Essentially D&D where the Manifold community controls one character and help it navigate through a story with a daily action to decide with free response markets each day. David and Ian worked on this concept for the hackathon we hosted yesterday and it was received well!
I'm not entirely convinced having some markets that are more adjacent to polls is necessarily something to avoid, especially if it is acknowledge it is a game. That being said we would have to be careful to avoid a situation where users bandwagon on the top response as all they care about is making a small % of M$ as opposed to caring about the game. We were thinking of creating a system where the top 3 responses are taken and one is chosen at random weighted to their probabilities. If we did want to make it very "prediction centric" we could make it so that the story is pre-written and users have to guess what event will happen or how the character will act, but I think this is less compelling than users actually impacting the story.
A big multiple-choice market where every square is an option. Every once in a while, the square with the highest bet gets revealed. If it's safe, the people who bet for it get a small payout. If it's a mine, everyone else who participated gets the payout instead and the game ends.
Free-response market where each option is a 5-letter word. Bettors put M$ on the word that they think will be the best 1st guess for the following day's Wordle. I'm not sure what scoring system would be best. Something like: +2 for green (right letter, right place) +1 for yellow (right letter, wrong place) +0 for red (wrong letter) would probably work. Potential alternatives to that scoring system: * market where red is penalized * market where yellow and red are both penalized Another possible version of this game could look like people betting on the worst possible guess, with something like: +4 for green +2 for yellow 0 for red where the goal is to get the lowest score possible.