The seller makes a claim: “this will sell for this price by this date.” They put down a deposit and right away announce a reward—a share of the sale price—for the people who put money in alongside them and help sell. These are the supporters. The people who bet against are the doubters. The argument is the lot’s advertising, and it springs up on its own: live, with money on the line, and the seller doesn’t pay for it.
Today a seller’s words are worth nothing: ads are bought, reviews are faked, and everyone writes “it’s about to take off.” A buyer can’t tell confidence from hot air.
Here confidence costs money. The seller publicly claims, “this will sell for this price by this date,” and puts down a deposit to back it up. Day after day, the deposit slowly drips away to the doubters until the sale happens. Anyone can join either side of the bet, and the bet itself advertises the lot—live, with money on the line, and at no cost to the seller.
When listing a lot, the seller announces up front a reward—a share of the sale price—for whoever helps
sell it. From there it’s simple:
· it sells—you pay the reward and the fee, like a referral fee; you get back what’s left of the deposit and 40–55% of what dripped;
· it doesn’t sell—the deposit has dripped away to the doubters, but no reward is paid at all: there was no sale;
· you can fail for two different reasons: the price is too high, or the item is weak—and the market will show which one it is;
· you can’t just add the reward on top of the price: supporters won’t back an overpriced lot.
You bet on the sale—so now it pays to help make it happen: tell people about it, bring in a buyer. If it sells, you get your share of the announced reward, by how much you put in × how long you held it (and 40–55% of the drop that has already dripped comes back; the platform keeps the rest). If it doesn’t, you get everything back except a small drop known in advance: the drop filters out people who don’t really believe. A small stake means a small share—exactly how small is worked out in “The honest truth about supporters.”
See an empty promise? Bet against it and take the dripped deposit if you turn out to be right. If you’re wrong, you lose time, a small fee and the interest you could have earned elsewhere—but not your own money.
Instead of star ratings and fake reviews—real money: you can see how much is staked on “it will sell” and how much against. Bets don’t lie: lying would cost too much.
In one sentence: a market where words are backed by money—and the argument over the price turns out to be the best advertising the item can get.
Stakes on both sides—“for” and “against”—sit in a shared safe for as long as they’re in (you can pull out at any moment). Nobody spends them, and they go back to their owners: doubters get theirs in full, supporters minus a small drop announced in advance. The seller’s deposit comes back like this: if it sells, what’s left of it comes back plus part of what dripped (40% for NFTs, 55% for expensive items, the platform keeps the rest); if it doesn’t, what dripped stays with the doubters for good. While the bet is live, all this money is tied up in the lot: everyone who put it in is an interested onlooker. Ordinary advertising buys a second of attention; here money holds the attention for as long as the stakes are in.
The reward is shared, and it’s split by the rule “how much you put in × how long you held it.” The doubters’ pot is split by the same rule. Say you put $10,000 on a Lamborghini selling while the “for” side totals only $60,000: your share of the $15,000 reward is $2,500. Then a crowd rushes in and the total grows to $600,000—now your share is $250. The early and the patient get the most; early entrants earn outsized returns, and that draws the crowd in on its own; as the crowd arrives, the edge fades. Once the market settles, both sides earn ordinary market interest—but that whole crowd has money riding on the lot. That’s exactly what we want.
Sets the price and the deadline, puts down a deposit and announces the reward up front—a share of the sale price for supporters. If it sells, the seller gets back what’s left of the deposit plus 40–55% of what dripped, and pays the reward out of the price. If it doesn’t, the deposit has dripped away to the doubters, and no reward is paid: there’s nothing to pay it from.
Puts their own money into the shared “it will sell” pool—and isn’t just allowed to promote the lot, but has every reason to. If it sells, supporters split the announced reward by how much they put in × how long they held it, and 40–55% of the drop that has dripped comes back to them (the platform keeps the rest). If it doesn’t, they get everything back except the drop known in advance (a small share of the stake that goes to the doubters) and the entry fee.
Bets that the sale WON’T happen: freezes money against the lot. If they’re right, they take what has dripped. If they’re wrong, they get the stake back in full and lose only the entry fee and time: while the stake is in, the money is frozen. They can leave at any moment—the stake comes back right away, and the share of the pot they earned before leaving stays theirs if the lot never sells. That’s the price of their vote.
Earns either way, but not equally. If the sale fails—only fees. If it sells—also the part of what dripped that didn’t go back to its owners (45–60%). So a sale pays it better: in the Lamborghini example below, about $15,900 on a sale versus $3,100 on a failure; in the demo lot, 0.156 ETH versus 0.018 ETH. That’s by design: the platform earns when things sell, so it has a reason to bring in buyers. And it can’t confirm a sale by itself: in the first version, the contract does that.
Small stakes from the platform itself on both sides of a new lot—so the game gets going even before any live players arrive. Winnings go to the fund for participation points (rewards for early participants); losses are the platform’s cost of getting things started. A one-way valve, and it applies only to seed stakes: they can lose money to the market, but whatever they win never reaches the platform’s pocket. The platform’s income is a separate thing: its stated fees and, on a sale, the part of what dripped that it keeps (see the card above).
The person who actually buys the asset. If they show up before the deadline, that’s the moment of truth: “sold.”
Both sides can argue out loud—and both back their words with money. Supporters praise the lot so it sells. Doubters, for their part, want more supporters to show up: every supporter’s stake feeds a drop into the doubters’ pot. Trashing the lot from day one would scare the supporters off, and the pot would shrink. So doubters usually stay quiet and speak up closer to the deadline. This argument is free advertising for the lot.
The seller has three separate money streams. Don’t mix them up—each one has its own job. On top of that, the platform hands out points—but those aren’t the seller’s money (more on them below).
Drips bit by bit—continuously, at a rate known in advance, for as long as the lot is up. The whole stream goes into the doubters’ shared pot; the seed stake makes sure the pot is never empty.
Announced when the lot is listed, paid only if it sells, out of the money from the sale. Split among the supporters: whoever got in earlier and held longer gets more.
A percentage of the sale + a small cut of everything that dripped from the deposit. On a sale, the platform also keeps the part of what dripped that doesn’t go back.
Points reward participation. The platform itself gives them to early participants to get the market going. There is no token, and there won’t be one until the mechanics are proven. Three rules: the mechanics have to work without points; points are awarded for how much you put in × how long you held it, not for clicks; and they’re handed out in proportion to the fees paid, so pumping up fake activity doesn’t pay: whoever farms points pays for their own prize.
The three streams form a closed system: whatever one side loses, another gains. No money is printed out of thin air.
The seller has announced the price, the deadline and the supporter reward; the deposit is locked in the contract.
The platform puts a little on both sides—the deposit starts dripping, and the market is alive from the first minute.
Doubters freeze money against the lot, supporters bet on it and bring in buyers. You can join and leave at any moment: your stake comes back right away when you leave.
The rate is known in advance. The stream is split among the doubters by how much they put in × how long they held it—and the market’s distrust shows in the ratios, not the raw amount: how much is staked against per dollar in the pot, and how “against” compares with “for.” The raw amount against grows with the deposit itself, since a bigger deposit draws more doubters.
Either a buyer shows up before the deadline or nobody does. There’s no third option.
The key to honesty: both sides of the bet risk time, fees and earnings; supporters also risk a drop known in advance. Only one person risks the deposit itself: the seller, backing their own claim. That’s what makes this a market that holds sellers to their word.
What each side pays and what it gets. All five examples were calculated by our simulation engine under the current rules, each with the settings for its type of item. In the other four examples, the sale happens 60% of the way to the deadline.
This is how it would work for real-world things that aren’t on a blockchain. The first version covers NFTs and tokens only (for why, see “Who confirms the sale”).
The dealer is sure: “It’ll go for three hundred grand within a month and a half, tops.” Too sure, a lot of people think.
How much of this goes to the blogger (about $50) and to a large supporter—the numbers are in “The honest truth about supporters.”
The developer: “This lot will go at the list price within a quarter.” The market is sluggish, and there are plenty of doubters.
The artist: “This piece will go for ten grand within a week.” Crypto Twitter is already placing bets.
A project sells spots on its presale allowlist (whitelist). Lots of people want in—but there are lots of scam projects around, too.
The band promises: “We’ll pack the venue—all 500 tickets will be gone within a month.” A local promoter smirks: “Nobody in this town knows who you are.”
A dollar spent on ordinary advertising freezes nothing around the product. Here everyone who placed a stake is an interested onlooker until the deadline.
The money that gathers around a bet can be many times what the seller spends—that’s a ceiling, not a forecast: how much actually comes is up to the market.
The model points to one pattern: a lot with a short deadline draws more money per dollar the seller spends than a long one. But the first lots will gather less than the ceiling—the market starts out empty, and it will have to be kick-started with seed stakes and points.
All the platform’s settings in one place. The lot and rule sliders are next to the chart; the controls under the chart change the chart itself. It’s a simplified model for getting a feel for what depends on what; its numbers are illustrative, not real.
The lot price is fixed: 10,000 coins. Everything else is a share of it, so the zeros don’t change the picture.
bp means basis point, a hundredth of a percent: 10 bp = 0.1% of the stake.
The market base rate is the yearly interest the same money would earn almost risk-free somewhere else (in a savings account, say). Doubters’ earnings are measured against it.
| Time passed | Fast start | Linear | Final sprint |
|---|
The four rows always add up to zero: no money is printed, it only changes hands. Left out: the lot price (a transfer from buyer to seller) and the stakes themselves—doubters get theirs back in full, supporters minus the drop (it shows in their row). Points aren’t modeled.
Doubters’ stakes come back in full; supporters’ stakes come back minus a small drop announced in advance (it filters out people who don’t really believe). Only one person takes a big risk: the one who made the promise.
Want out of a promise that’s failing? There’s an honest “I give up” button with a clear price: withdraw the lot early and pay the penalty. Faking a sale to yourself costs more: the fee, what the platform keeps and the reward to the real supporters are all paid for real. We don’t catch cheaters—we send them the bill.
On ordinary prediction markets, influencing the outcome is forbidden. Here, promoting the lot is the game: a supporter brings the sale closer by doing the work, a doubter stress-tests the promise. It’s work, not roulette.
We don’t catch cheaters here—we send them the bill. Before writing any contracts, we set simulated cheaters on the mechanics: they know everything and pick the most profitable trick. We found nine ways to cheat. Eight have been calculated: usually cheating costs more than the legitimate path, and where a gap remains, we say so plainly (as with list-and-cancel below). One (drop outsourcing—a service pays the drop for supporters and splits the reward) is still open: nobody in the industry has a ready fix. Here are the three main ones:
The seller “sells” the lot to themselves through a second wallet, to fake a success and save the deposit that’s dripping away.
Defense: the fee + the part of what dripped that doesn’t come back. The margin—how much more a fake costs than giving up early—doesn’t depend on the number of doubters and gets thinner toward the deadline (in the demo lot, about 8.5% of the price on day one and about 3.7% at the very end), but it never disappears: giving up early is cheaper than a fake on any day.
The same seller bets “for” themselves from other wallets, to take the reward back when a real sale happens.
Defense: reward shares are weighted by how much you put in × how long you held it, and a planted stake itself loses the part of the drop that doesn’t come back—and the reward is the seller’s own money anyway: a fake can’t take anyone else’s. To be clear: this is a residual risk, but it has a ceiling. The cheater shortchanges their real supporters on part of the reward—in the demo lot, no more than about 3.6% of the price; beyond that, the drop on the planted stakes eats the profit. The fix: before the reward is paid out, check whether the stakes came from the seller’s wallets, plus reputation (more in the FAQ).
The seller bets “against” themselves, keeps listing and pulling lots, and pockets the drops of the honest supporters they lured in. The cancellation compensation goes from the seller to themselves.
Defense: without bait, this game costs the cheater about nine times more than the harm it does. Still, a gap remains here. If real supporters came to the lot, the cheater makes about 0.4–1% of the price per cycle. The fix: everyone can see the history of cancellations, and people stop trusting a serial canceler; plus a small fee for listing a lot—it has been designed but isn’t switched on yet (the author’s call).
All nine ways—with the numbers “your result versus the honest path”—are in the demo lot: try to cheat.
Every setting in the mechanics, in one table—who sets it and what it means.
| Setting | In plain words | Who sets it |
|---|---|---|
| Price and deadline | The claimed price and the deadline by which the lot should sell | seller |
| Deposit | The seller’s money behind the promise; drips to the doubters until the lot sells | seller |
| Reward | A share of the sale price for supporters, announced when the lot is listed | seller, within the rules |
| Drip shape | How fast and along what curve the deposit drips; known in advance | platform rule |
| Drop | The share of a supporter’s stake that drips over the lot’s full run: on a failure it goes to the doubters, on a sale part of it comes back (see the next row) | platform rule |
| Refund on sale | What share of what dripped goes back to its owners if the sale happens | platform rule |
| Sale fee | A percentage of the price | platform income |
| Drip fee | The platform’s cut of everything that dripped—both the deposit and the supporters’ drops. Taken once, when it lands in the pot; after that the pot is clean, with no double charge | platform income |
| Entry fees | Fees on both sides’ stakes—fractions of a percent. If you guess wrong, this is what you lose | platform income |
| Doubter penalty | Paid by a doubter if the sale happens. Zero by default; if it’s ever switched on, the money only gets burned | backup setting |
| Points | How many points per unit of fees paid | start-up boost, fades out |
| Seed stakes | The size of the platform’s stakes for the doubters and for the supporters | kick-starts an empty market |