Onchain Atlas

Conditional Tokens

Tokens that only become real, tradeable value once a specific real-world question is answered — one for each possible answer.

Also called: outcome tokens · conditional token framework · CTF

What it is

Conditional tokens split an underlying asset (like a stablecoin) into a set of outcome tokens, one per possible answer to some future question ("Will X happen by date Y?"). Holding all outcome tokens together is worth the same as the original asset; holding just one outcome token is a bet that specific outcome will occur, because only that token can be redeemed for the full value once the question resolves.

How it works

  1. A user deposits collateral (e.g., USDC) into a conditional tokens contract and specifies a "condition" — a question with a fixed set of mutually exclusive outcomes, tied to an oracle that will eventually report the answer.
  2. The contract mints one outcome token per possible answer, each fully backed by the deposited collateral, and gives all of them to the depositor. At this point the set is fungible with the original collateral (split it, and you can always merge it back).
  3. Outcome tokens are transferable ERC-1155-style assets, so they can be sold individually on an exchange or AMM — this is what creates a prediction market or hedging instrument, since the price of an outcome token reflects the market's implied probability of that outcome.
  4. Conditions can be nested: an outcome token from one condition can itself be used as collateral for a second condition, letting builders construct combinatorial markets (e.g., "outcome A AND outcome B").
  5. When the oracle reports the resolution, the contract marks which outcome was correct.
  6. Holders of the winning outcome token redeem it 1:1 for the underlying collateral; all other outcome tokens become worthless.

Why designers use it

  • Turns any resolvable question into a tradeable, fully-collateralized instrument without needing a centralized bookmaker to hold funds or set odds.
  • Because the full set is always redeemable for the collateral, there's no counterparty risk beyond the oracle itself — the contract can't print money it doesn't have.
  • Composable: outcome tokens are ordinary tokens, so they can be listed on AMMs, used as loan collateral, or combined into more complex conditional structures.
  • Enables permissionless market creation — anyone can define a new condition and let the market form prices around it.

Failure modes

  • The whole system inherits the oracle's trust assumptions: if the oracle reports incorrectly or is manipulated, all outcome tokens settle wrongly regardless of what actually happened.
  • Ambiguous question wording ("will X happen by date Y" with no precise resolution source) leads to disputed resolutions and split communities, as seen in real prediction markets during contested events.
  • Low liquidity in a given outcome token means its price is a weak signal of true probability, undermining the "market as forecast" pitch.
  • Combinatorial/nested conditions can create outcome tokens so thin that no real market exists, leaving capital locked with no practical way to exit before resolution.
  • If merge/split accounting has an edge case bug, a user's full-set collateral guarantee can break, exposing the protocol to insolvency.

What to check before using it

  • Who or what resolves the condition, and what happens if the oracle is late, wrong, or captured?
  • Is there a dispute or challenge window before a resolution is treated as final?
  • What's the expected liquidity for each outcome token — is there an AMM or market maker committed to quoting it?
  • Are questions worded precisely enough that a reasonable oracle can't disagree about the answer?
  • Does the collateral asset itself carry risk (depeg, freeze) that would undermine the "fully backed" guarantee?

Experiments that used it · 3

Shown oldest first, so you can watch the design evolve.

Gnosis Conditional Tokens
A generic, oracle-agnostic ERC-1155 framework for tokenizing outcomes of future events, enabling combinatorial prediction markets and becoming the settlement primitive under Omen and Polymarket.
2019 major success
Polymarket
The largest onchain prediction market, built on Polygon using Gnosis Conditional Tokens, a hybrid off-chain order book, and UMA's optimistic oracle for resolution.
2020 major success
MetaDAO
The first live onchain futarchy, where prediction markets on a token's price — not token-holder votes — decide governance proposals, later pivoting into a futarchy-governed token launchpad on Solana.
2023 ongoing