What Happens if a Prediction Market Expires With No Trades?

27 September 2026

If a prediction market expires with no trades, the outcome is straightforward: all open interest is zero because none exists, and the market simply closes without any payouts, losses, or settlements. No one wins or loses because no one held any shares. The market creator, however, still faces the costs they committed to upfront - namely, the market creation bond - and the collateral they deposited for liquidity remains theirs to withdraw.

This scenario is rare in active markets but entirely possible for obscure or poorly promoted ones. Understanding what does not happen is as important as what does.

The empty order book

Every prediction market starts with a creator who defines the event, sets the resolution criteria, and deposits collateral (often USDC or DAI) to seed the initial liquidity pool. That pool is what allows others to trade against. If no one ever buys or sells a single share, the pool remains untouched. The automated market maker (AMM) formula - typically a constant product curve like x * y = k - still holds, but with no trades, the price of each share stays exactly where the creator set it.

The market still runs its course. The resolution date arrives. The oracle - whether UMA, Kleros, or a custom resolver - submits the outcome based on real-world data. The market resolves to "Yes" or "No" (or a range, if it's a scalar market). But because no one holds any shares, there is nothing to redeem. The resolution is a formality.

What the Creator Loses and Gains

The market creator's primary financial exposure is the creation bond. This is a deposit paid at market creation to deter spam and frivolous markets. It is typically returned once the market resolves, provided the creator did not try to game the system. But "returned" does not mean "profitable." The bond is locked for the market's entire duration, which could be weeks or months. That's an opportunity cost.

The liquidity the creator deposited is separate. That collateral sits in the AMM pool. With no trades, the pool's value is unchanged at expiry. The creator can withdraw the entire amount - minus any trading fees that the AMM would have accrued, which is zero. So the net effect is: the creator is out the time value of their bond, but not the principal.

There is no penalty for an empty market beyond the bond. The protocol does not fine a market for low activity. It simply resolves and closes.

The settlement process step by step

If you are a market creator or a curious observer, here is what actually happens on-chain when an untraded market expires:

1. The resolution time passes. The market's designated end time arrives. No further trades are possible. The order book - if it was ever populated - is now frozen.
2. The oracle submits the outcome. A dispute window opens. For optimistic oracles like UMA, anyone can challenge the proposed outcome within a set period. For Kleros, a jury may be involved. In an untraded market, this step is usually uncontested, but the protocol still follows its standard procedure.
3. The market resolves. The outcome is written to the smart contract. For a binary market, the result is 1 or 0. For a scalar market, it's a number.
4. The creator claims their bond. After the dispute window closes, the creator calls a function to reclaim the creation bond. This is not automatic; it requires a transaction.
5. The creator withdraws liquidity. The AMM pool is still funded with the creator's initial deposit. The creator removes it. If the pool had any other liquidity providers (which, by assumption, it does not, because no one traded), they would claim their share proportionally.

In practice, step 4 and 5 are often combined into a single transaction. The protocol may also allow anyone to trigger the bond return on the creator's behalf, though this is an implementation detail that varies by platform.

Why this seems weird (and why it isn't)

Newcomers often expect an untraded market to "expire worthless" or to "auto-resolve to No." Neither is true. A market with no trades has no counterparties. The concept of "winning" or "losing" requires someone to hold a position. With zero shares in circulation, there is no one to pay.

This is different from a market where someone buys a share and then fails to sell before expiry. In that case, the holder's position is settled against the resolution outcome. But if no one ever buys, the market is a shell.

The lack of trades also means the market's price feed - the thing that shows "Yes: $0.42" - is meaningless. That price was set by the creator's initial parameters, not by any market activity. Anyone looking at an untraded market should ignore the displayed price entirely.

Edge cases and gotchas

A few situations complicate the "no trades" scenario:

• Partial fills on limit orders. If a trader placed a limit order but it never matched, the order sits in the book. At expiry, the order is cancelled. The trader's collateral is returned. No settlement occurs because the order was never filled.
• Liquidity provision by third parties. If the creator seeds the pool, but a third party also adds liquidity, that third party has a claim on the pool. Even if no one trades, that liquidity provider can withdraw their share at expiry. They may lose the spread, but they don't lose principal.
• Market with a single trade. The question assumes zero trades. But if even one share changes hands, the entire dynamic shifts. The holder now has a claim that settles against the oracle's resolution. The market creator's liquidity is partially consumed by that payout, and the AMM's invariant is broken.

The real risk is neglect, not loss

For a trader, an untraded market presents no risk because there is no exposure. For a creator, the risk is opportunity cost - the bond and liquidity are locked, earning nothing, for the market's lifetime. That's it.

The more interesting question is why someone would create a market that no one trades. The usual reasons are: they misjudged interest, they created it too far in advance, or they set a resolution criterion that was too narrow or too niche. None of these are protocol failures. They're market failures, in the most literal sense.

In the end, an untraded prediction market is a non-event. The code runs, the oracle resolves, the bond is reclaimed, and the collateral returns. Everything else is silence.

Back to prediction markets