Let's play a variation on poker! Like all the decent games of poker, this will be a lot about leveraging luck with strategy and mind games; a perfect match for a place about predictions imo.
We're playing a variation based on Texas Hold 'Em specially designed for working on Manifold over a series of markets, but everyone can participate and bet even if they're not given a hand.
Here's how it works:
This question is the buy-in for the hand. The bettors with the top 7 largest YES positions at close get a personal hand. Bots and alts are allowed, but I'll need a way to contact you to give you your hand, such as discord. I will be trading in this question, and any profit I make (if any) will turn into a bonus prize for the winner of the hand.
After this closes and I deal out the hands, I'll draw a card. If it's black, this question resolves YES; if it's red, it resolves NO.
Players who get a hand are welcome to share it, or team up, bluff, obfuscate, and misdirect, whether on the questions, over discord, or wherever they see fit. They're also welcome to bet for or against their opponents!
Once everyone has their hand, I'll make a general question about "who will win" that will track the status of the game, and then rapid (~1 day) questions for each of the four rounds of betting.
Each player must commit a certain amount of mana on YES on themselves each round to advance to the next stage, as indicated by the round's trading history, otherwise they fold. Their commitment is the net of YES mana spent minus NO mana spent on themselves. This minimum equals half the lowest buy-in position that got into the game. (I'll announce it once the game starts)
If any player commits double or more than the minimum in any round, the minimum is doubled in all future rounds. i.e.: if the minimum mana to play is 20 and one or more people commit 40+, all future rounds have a minimum of 40, unless they get doubled further.
The Audience Hand works exactly like personal hands, except it:
is publicly visible
doesn't have to buy in
must have twice the minimum, using bets from all bettors, to advance
The round questions resolve at the end of the game to the winner(s), equally.
If two or more players stick it out to the end I'll reveal their hands, announce the winner, and resolve all the remaining questions accordingly.
If for whatever reason every player folds at any stage, the win goes to the player who held the highest bet on themselves in the previous round.
This hand is the first in a short tournament series. There are several ways to score points in the tournament (even if you don't get a hand), but your best chance is to get into the game.
As you might have noticed, the rules of this special variant allow for mismatched bets by players, so feel free to jump in even if you're intimidated by a whale in the water! They can do their thing, profit by playing market dynamics, and you can still come out on top :) Admins and Manifold staff also have no inherent advantage, so don't be shy!
As I'll know all of the cards, I won't be betting or playing, and I cannot be bribed to reveal other players' information or tilt in your favor. My house cut is simply the trader bonuses on any questions I make for this, though I'll accept tips for my role as dealer if you're so inclined.
Of course, side markets are welcome. Feel free to add them to the Manifold Plays Poker group and comment them here for shared visibility, and they'll count for the tournament!
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ38 | |
| 2 | Ṁ37 | |
| 3 | Ṁ14 | |
| 4 | Ṁ10 | |
| 5 | Ṁ9 |
How about some "Bribe Pat" options for Hand 5 to build money for the pot and to give betters more information with which to work? E.g. (numbers arbitrary):
Pay 30 mana to trade out one card at the deal stage (random replacement)
Pay 60 mana to trade out one card at the flop stage (random replacement)
Pay 50 mana to privately see 13 cards in the not-to-be-used pile (which gives you information about what might / might not be coming in future rounds)
Pay 50 mana to privately learn the total number of cards in each suit across all eight dealt hands
Pay 100 mana to resurrect a folded hand in the round after you folded (works only prior to the river)
I'd suggest that everyone learns that a specific bribe was made, but not the identity of the bribe-giver except for resurrection of course.
@Jason omg this got buried in my notifications. but luckily I have direct phone pings on for some types of comments, which sometimes works after a big delay.
anyway, these are some interesting ideas! I think we could even try doing a few with subsidies, which naturally grows the pot without having to managram me or dm a manalink. I'll think about how to add something like this in for Hand 5 when I go to design it.
@Stralor despite yet again not earning a prize pool, I think this is a decent way of handling these questions. I'm open to other ideas, though! lmk
@Stralor the previous one, where resolution was determined, seemed to make more sense to me. What do you see as the advantages of this one?
@Fion not a big fan of the whole market sitting at 1-10%, not much margin. I figure this way sometimes I build a decent prize pool or sometimes y'all get a bonus for playing
@Stralor I am surprised this was not more popular. Either entrants make a profit or there is a small cost of getting into game but there is a ~7*larger prize prize fund that might be won.
i.e. surprised you had to sell at as low as 52% to get participants. Extending deadline rather than selling that low makes more sense? Maybe buy up to 59% when market opens and set sell orders at 60% 62% 64% ... ?
@ChristopherRandles 🤷 maybe people have had enough poker? maybe the buy-in is (rightfully) the least important part to players so they don't engage strongly if there's a good chance of no profit? idk! I'll keep experimenting until I'm done with this series or everyone else seems to be
@Stralor but that prob would help with guaranteeing a prize pool more often. I have dramatically new ideas for when added-answer multi binaries are released, but might loop back to this idea just like I might loop back to the Pivot method from Hand 2
@Stralor I think you should win enough money, in expectancy, in the buy-in phase to justify a meaningful prize pool. How about resolve-to-20%, but you put in a massive NO limit order at 21% so you can capture most of the amounts bought in for the prize pool?
Pivot method had viability for earlier stages, but allowing a single user to set the pivot amount in the final round was too whalish even though no one tried to whale it.
@Stralor If people want to make profits on all (or as many as possible) markets traded. 7 people who send you largest total in managrams labelled 'hand x buy in' (minimum 10 mana?) get a hand and this becomes prize pool. If you don't get 7 extend deadline. This makes it a gift rather than a trading loss? Would entrants prefer that?
Main question and deal are up
https://manifold.markets/Stralor/manifold-plays-poker-who-will-win-h-46e529ba522a?r=U3RyYWxvcg
@Stralor here we go (don't panic, that's the audience hand. I'll make the markets, screenshot the hands, and send everything out imminently)
Start of 3 game poker tournament call. Might as well try and get some more traders to spice it up a little.
Differences this hand:
Not chaos mode
Minimum bet thresholds based on YES mana spent (not shares)
No Pivot
Irreversible betting(might not work as I hope, open to tinkering)Increasing minimums
@Stralor "Only players' YES bets on themselves count for advancement and can't be retracted by equal NO bets or bets on other players." Does that mean if I bet M100 YES and then immediately M100 NO (for a net position of 0) I'll still advance?
@Stralor maybe it would help to explain what this rule is intended to accomplish so we can tinker with wording? I think the problem is coming mainly from an unclear use of "retracted."
@A yeah, that's how it works at the moment but I think it's actually a terrible idea. maybe we should stick with YES and NO mana counting?
@Stralor Could we require a net YES position and determine the cost basis of the net position to assess continuation? Agree that basing retention on gross YES spend on yourself doesn't make any sense if the point is to require people to put up some mana in order to stay in.
@Jason agreed. the problem with using position in multi binaries is that it takes a ton of work. we tried it, and until Manifold releases UI upgrades that allow you to see other traders' positions I just don't think it's worth it.
@Stralor alright for now I'll adjust it to net YES mana spent. doesn't make sense to break the game for a hand