Zeitgeist answers a different question from the regulated exchanges on this list. Kalshi and Polymarket US are licensed venues where the product is a dollar-settled contract and the licence is the moat. Zeitgeist is a public blockchain where the product is a token, the rulebook is code, and the ambition is to be the layer other people build prediction markets on — its own documentation describes the chain as “designed and optimized to be used as an SDK for anyone looking to implement their own prediction market technology”.
That makes it a different kind of listing to compare, and a different kind of risk to take. There is no operator to hold accountable, no supervisor to escalate to, and no fiat on-ramp documented in its own materials. Everything below is taken from zeitgeist.pm and docs.zeitgeist.pm; where those sources are silent, this page says so rather than filling the gap.
A Layer 1 built on Substrate
Zeitgeist runs its own Layer 1 chain on Substrate, the framework behind Polkadot and Kusama. It is not a contract deployed on Ethereum and not an app renting space on someone else's rollup: the markets, the liquidity pools behind them, the dispute court and the governance machinery are runtime modules on Zeitgeist's own chain. The documentation names them directly — zrml-prediction-markets, zrml-neo-swaps, zrml-court and zrml-futarchy — which matters when you compare it with a smart-contract protocol, because fees and upgrades are decided at the chain level rather than by contract migration.
The project's own framing is decentralisation-first: “No single entity controls our network, and its distributed governance mechanisms ensure it stays that way.” As the governance detail further down shows, that claim carries qualifications built into the runtime.
How a Zeitgeist position actually settles
This is the part that catches people arriving from a dollar venue. The base asset is ZTG, the chain's native token, and outcome tokens redeem for 1 ZTG per winning share — not for one dollar. Trading fees are paid in ZTG. The genesis supply was 100 million ZTG, counted on-chain in Pennocks at 10,000,000,000 to the token. The docs note that selected foreign assets can also serve as collateral, but ZTG is the default and the only asset accepted for the bonded actions described below.
So a winning trade is not a dollar profit. Your payout is denominated in a volatile asset, and your realised return is the market call plus whatever ZTG did between entry and redemption. A market trading at 0.60 here and 0.60 on a dollar venue is not the same payoff described two ways; it is two different payoffs.
Categorical and scalar markets
Two shapes exist. A categorical market lists discrete outcomes — Yes/No, or every runner in a race — and pays the winning token in full while the rest go to zero. Zeitgeist's guidance for anything beyond a yes/no question is to add a catch-all option, and the docs point to the Kusama Derby as the cautionary tale: when the defined outcomes all missed, the catch-all token “None” was what actually resolved.
A scalar market bets on a number inside a range set at creation, using Long and Short tokens rather than categories, and redeems them proportionally. With a range of $1,000–$1,200 that resolves at $1,180, Long pays 0.90 and Short pays 0.10; resolve above the upper bound and Long pays 1.00 while Short is worthless. This is the shape futarchy runs on, and the shape people mean when they say a prediction market can forecast a price rather than an event.
Creating a market is a bonded, gated action
Anyone can create a market, but not for free, and the documentation is specific about the two routes.
| Route | Bond | What happens next |
|---|---|---|
| Advised | 200 ZTG | The Advisory Committee decides whether the market is valid. Until it rules, no liquidity pool may be deployed and no complete sets bought or sold. A rejected market forfeits the bond. |
| Permissionless | 1,000 ZTG | No approval needed and the market can trade immediately. Zeitgeist reserves the right to remove it later, and says it will where it expects 10x more harm than benefit. |
Bonds are returned once a market resolves cleanly. A separate oracle bond is posted by whoever reports the outcome, and it is slashed if the report is not faithful. The docs flag the obvious trap: the market creator names the oracle and funds that bond, so naming an address that never reports burns the creator's stake while the unwitting oracle walks away untouched.
The published rejection rules are candid about why the gate exists. Markets are refused for vague questions, missing resolution criteria, outcomes that ignore major possibilities, and start or end times that make no sense — on the reasoning that “if it has any chance of confusing people or resolving ambiguously, we will not approve it”, and because a market that resolves too soon cannot be taken back.
Resolution: oracle, grace period, then disputes
Every market runs the same sequence. It opens at its start block or timestamp, closes at the end, then sits in a grace period during which all activity halts so trading stops before anyone reports. The oracle then has a fixed reporting window. Miss it and the creator's stake is slashed while any address can step in and report as an outsider; the market moves from closed to reported, not yet resolved.
That report is provisional. A dispute window follows, and the mechanism used inside it is chosen when the market is created — either Authorized, where the Advisory Committee resolves it, or the on-chain court. Only when the window closes without a successful challenge does the outcome become final and redeemable.
The court: commit, reveal, appeal
The court is Zeitgeist's most distinctive design and its most ambitious claim. It is a stake-weighted plurality mechanism built on the Schelling point: jurors vote as encrypted hashes during a vote phase, then reveal during an aggregation phase so every raw vote can be matched to the hash it came from, and the outcome with the most revealed votes wins. Zeitgeist credits Kleros as the model.
Jurors and delegators lock ZTG to take part, and weight decides who gets drawn into a case. Three appeals are allowed; at the fourth the case leaves the court entirely and becomes a global dispute, in which every ZTG holder can lock funds behind an outcome and the heaviest side wins. The economics are hard-coded rather than discretionary: a minimum juror stake of 500 ZTG, a first-round draw of 31 times that (15,500 ZTG), and by the final appeal round 255 draw weights — roughly 127,500 ZTG — pulled from the pool. Jurors who vote against the plurality forfeit a multiple of the minimum stake, and that money is redistributed to the jurors and delegators who backed the winning side.
Two details are worth reading before treating a court-governed market as safer than a straightforward one. The participant pool is capped (MaxCourtParticipants), so the lowest-staked participant is displaced when a larger one joins, and exiting takes an inflation-period lock before funds return. Decentralised resolution is real here, but it is slow, capital-intensive, and depends on enough honest stake showing up in the draw.
Futarchy: the part nothing else on this list does
Zeitgeist's headline governance idea is futarchy — Robin Hanson's “vote on values, but bet on beliefs”. The welfare measure is set by a vote; policy is then decided by markets, because prices aggregate what informed traders actually believe will happen. The chain implements this in the zrml-futarchy pallet, added in runtime v0.6.0.
The mechanism is narrower than the marketing suggests. A proposal is submitted with two markets — one on whether the proposal will be enacted, one on the welfare measure at a future time — combined through a combinatorial pool into four positions. If the long-if-enacted position prices above long-if-not-enacted, the proposal is judged to improve welfare. The trigger is deliberately conservative: only root can call the submission extrinsic, so a futarchy vote must first clear an ordinary token-holder democracy proposal to get there at all.
Liquidity comes from pools, not market makers
There is no market maker quoting both sides; depth comes from liquidity pools, and the providers earn the fees paid by swappers while carrying the risk of being left holding losing outcome tokens. The docs put that trade-off bluntly: as a liquidity provider on a prediction market, “you are essentially betting against the informants' ability to predict future outcomes”.
Two designs appear in the documentation. Default pricing uses Zeitgeist's own Rikiddo scoring rule, a liquidity-sensitive automated market maker built on the Othman–Sandholm–Pennock–Reeves model, and the docs note that only one scoring rule is currently available. Separately, parimutuel markets use a “losers pay winners” pot that needs no liquidity at all: every bet goes into a shared pot and the winners split it in proportion to their stake. It suits short-lived casual markets, works only for categorical markets, and the docs list its drawbacks rather than hiding them.
Who it suits
Zeitgeist is the pick for someone who wants the mechanism itself rather than a product around it: conditional and combinatorial bets that span markets, exposure denominated in a chain token instead of dollars, and resolution decided by staked strangers rather than a company. The combinatorial work is inherited from Gnosis's conditional tokens design, and it lets you bet on a contingency — that a candidate wins if they are nominated — rather than a single event. For a trader whose main needs are liquidity, tight spreads and a known counterparty, the venues with dollar settlement and published volume are the better fit, and Zeitgeist's own documentation does not pretend otherwise.
What to check before you deposit
Four things. First, that the market you want is deep enough to fill your size, because all depth is user-provided and slippage in a thin pool is unforgiving. Second, which dispute mechanism that specific market uses, since Authorized and court-resolved markets behave very differently once an outcome is contested. Third, that you can source ZTG and get it back out, because no fiat rail appears anywhere in Zeitgeist's own materials. And fourth, read the resolution criteria and the catch-all option before trading: on a chain where a market can be removed after the fact, and where a badly written question resolves to whatever the criteria literally say, the market text is the contract.





































