Perpetual futures vs prediction markets: payoff and liquidation

Where the money comes from

A perpetual futures contract, or perp, tracks an underlying asset and has no expiry date. You post collateral, described in the Polymarket Perps documentation as pUSD, and your position gains or loses as the underlying price moves. A prediction market contract works differently: it represents a specific event with stated resolution criteria, and an ordinary winning binary share settles at 1 USD equivalent while a losing share settles at 0. Prices on binary shares are commonly read as implied probabilities, subject to the market's mechanics and costs.

So the perp question is where the price goes before you close, and the prediction market question is what the resolution decides. A binary share cannot be liquidated because you paid its price in full; it can end at 0. A perp can be closed at a loss at any time and can be closed for you by the liquidation engine, which the rest of this comparison takes apart.

Margin is not the same as notional

In a perp, initial margin is the collateral that lets you open exposure, and maintenance margin is the minimum equity required to keep the position open. Notional is the size of the exposure itself. At 100 margin against 1,000 notional, you control 10 times your collateral, which is where the leverage sits. Equity moves with profit, loss, funding and fees, and the margin mode you pick changes how much collateral is available. Leverage increases exposure relative to the posted margin, so the calculation must include the resulting liquidation risk.

A prediction market position does not have this split. You pay the share price and that is your capital at risk; the worst ordinary case is the share going to 0. There is no maintenance threshold to breach and no engine that closes you out for falling below it. That single structural difference drives most of the payoff and liquidation gap between the 2 products.

The 10% adverse move, worked twice

Take a hypothetical perp position: 100 margin, 1,000 notional, effective leverage of 10x. If the underlying moves 10% against you, the gross loss on notional is 1,000 times 10%, which is 100. Before fees or funding, that wipes the posted margin. The number is arithmetic, not a venue rule. Actual liquidation can occur before your margin is fully exhausted by gross price loss, because the engine reacts to maintenance requirements and mark price rather than to your original deposit running out. Fees and slippage can increase the loss. Funding can be paid or received, depending on your side and the applicable rate; include the signed payments in your calculation.

Now run the same adverse move on a binary prediction market share. A 10% drop in the share price before resolution is not a liquidation event. You can sell at that price if there is a buyer, or you can hold to resolution and receive 0 or 1. There is no funding charge and no margin call. The trade-off is symmetric: your maximum loss on the position is the price you paid, while a leveraged perp can be closed by the liquidation engine before your margin is fully consumed by the gross price move, depending on the venue's maintenance and mark rules.

Funding payments can be paid or received

Because a perp has no expiry, funding payments between the long and short sides tie it to the underlying reference. Direction, rate and interval come from current protocol parameters, so there is no fixed yield and no universal number to quote. If you hold across funding intervals, the accumulated payments are part of your holding cost and shift your effective entry and exit economics. An unchanged market price can still leave a different net result after funding paid or received.

A binary share held to resolution does not carry that drain. Trading costs and the spread affect what you can get before resolution, but the payoff at resolution is 0 or 1 as defined by the event. This is the clearest case where the 2 products diverge for a patient holder: the perp can be slowly consumed, while the binary position waits.

Fees use different bases, so headline rates mislead

Perp fees apply to trade notional and sit in tiers influenced by 30-day volume; entry, exit, funding and liquidation costs are separate items. Binary prediction market fees use a different structure, sometimes written as the share-price-dependent term C times rate times p times (1 - p). The point is not that one product is cheaper. The formulas are not comparable until you convert them to your own position size and holding period, and venue-specific numbers should be checked against the current schedule.

For a person comparing them, the useful habit is to build the total cost of the intended trade on each side: notional times fee tier times round trips, plus funding over the intended hold, against the price and spread you actually pay for the binary share. A low headline rate on a large notional can cost more than a higher-looking rate on a small one.

The differences side by side

The table contrasts the 2 products on the dimensions that decide payoff and liquidation. It is an illustrative summary of documented mechanics, not a ranking. Check fees against the current venue schedule for your volume tier, product and holding period.

What the example does not cover

The 100 margin and 1,000 notional example is a clean arithmetic demonstration, not a tested strategy. It ignores fees, funding, slippage and gap risk, and it assumes a linear 10% move against the position. It also says nothing about whether 10x is appropriate for anyone. The relation between a 10% move and liquidation depends on the venue's maintenance requirement, mark price source and risk controls at the moment of the move, and those can differ.

Funding parameters change over time, and no single rate applies across venues. A perp price is not a probability forecast and reading it as one is a mistake; probabilities belong to binary outcomes with resolution criteria. If you are sizing either product, compare the 2 on margin mode, maintenance requirement, mark price source, funding interval and fee tier for the specific venue, and keep in mind that a liquidation can arrive before the simple margin-arithmetic line.

To distinguish price windows and thresholds, see crypto prediction markets.

To examine losses and execution limits, see prediction-market risks.

To connect your estimate, the price and the costs, see expected value.