- Different questions. Perps: "where is this price going, continuously". Prediction markets: "will this event resolve true by this date".
- Perps has leverage up to 20x, hourly funding and forced liquidation. Prediction markets have none of the three.
- Maximum loss: Perps — your collateral, and it can arrive in minutes. Prediction markets — the price you paid for the shares.
- The decisive difference is time. In a prediction market you can be early and still collect. In Perps, being early gets you liquidated before the move arrives.
- Rule of thumb: if your thesis contains "eventually", it is a prediction market thesis. If it contains "before my stop", it can be a Perps thesis.
The structural comparison
| Polymarket Perps | Polymarket prediction markets | |
|---|---|---|
| Question answered | Where is the price going, continuously | Will a specific event resolve true by a date |
| Instrument | Perpetual futures contract tracking an underlying | Binary YES/NO shares in a defined event |
| Expiry | None — held until closed or liquidated | Fixed resolution date |
| Settlement | Continuous P&L; no resolution event | $1 or $0 at resolution |
| Leverage | Up to 20x (10x on many RWA) | None |
| Forced exit | Yes — on margin breach, at market | No — the position cannot be closed by the venue |
| Maximum loss | Your collateral; can arrive fast and with slippage | The price you paid for the shares |
| Ongoing cost | Funding, settled hourly | None while held |
| Cost on entry | Taker/maker fee on notional, per fill | Embedded in the spread; per-category taker fee |
| Being early | Fatal — liquidation does not wait for your thesis | Survivable — you hold to resolution |
| Position sizing | Notional must be derived from a risk budget | Stake is the maximum loss by construction |
| Natural use | Defined-stop directional trades, hedging, precise leverage | Event theses, asymmetric payoffs, anything needing time |
Loss profile against an adverse price move. The prediction market line stops falling — the loss is capped at the stake and time does not consume it. The Perps line keeps falling until it hits zero collateral, and it gets there on a move of a few percent.
Time is the difference that decides everything
Every other distinction in the table above is downstream of this one. In a prediction market, time is neutral. Your position pays nothing while it sits, nothing forces you out, and if the event resolves your way in nine months you collect exactly what an earlier entry would have collected — often more, if you bought cheap.
In Perps, time is charged. Funding is debited every hour, which reduces equity, which pulls the liquidation threshold closer. The position is not only betting on direction; it is betting on speed. A 20x position with a 2.5% liquidation buffer and no stop is a bet that the market does not move 2.5% against you before it moves your way — and at bitcoin's ordinary hourly volatility, that is a coin flip dressed as a view.
Can you be right and still lose? Yes, two different ways
| Failure mode | How it happens | Which product |
|---|---|---|
| Wrong, out of money | Market moves against you and you exit at a loss | Both |
| Right, but early | Liquidation fires before the move you predicted arrives | Perps only |
| Right, but funding ate it | Directional P&L smaller than accumulated hourly funding | Perps only |
| Right, but ADL closed it | Auto-deleveraging reduces your winning position in a violent move | Perps only |
| Wrong, but capped | You lose the stake and nothing more; the world moves on | Prediction markets only |
Three of those five are exclusive to Perps and all three can happen while your analysis was correct. That is not a flaw in the product — it is what leverage and a clearing mechanism do. It is a reason to know which product your edge actually depends on.
Matching the product to the thesis
| Your thesis sounds like | Fits | Why |
|---|---|---|
| "This event is underpriced at 62¢" | Prediction market | A probability estimate against a defined outcome, with a resolution date |
| "Bitcoin goes up this week" | Perps, if you have a stop | Directional, short horizon — but time is charged and liquidation is real |
| "The market will figure this out eventually" | Prediction market | The word "eventually" is disqualifying for any leveraged product |
| "This rally is overextended, I want to short into it" | Perps, if sized properly | You cannot short a prediction market the way you short an asset |
| "I want a 10x payoff on a 10% probability" | Prediction market | A 10¢ YES share is a natural 10x with a capped loss and no liquidation |
| "I want to hedge a spot position I already hold" | Theoretically Perps | Direction matters and precision matters — but only with capital you can lose |
| "I believe the price target will be hit by year end" | Prediction market | A milestone question with a date, not a continuous price bet |
What each one actually costs
| Cost | Perps | Prediction markets |
|---|---|---|
| Entry | Taker 0.04% or maker 0.0125% of notional at base tier | Per-category taker fee on shares × rate × price × (1 − price), plus spread |
| Holding | Funding every hour — 0.01%/h on $10k is $72 a month | Nothing |
| Exit | Another fee on notional, plus slippage; elevated fee rate if liquidating | Spread on the way out; no fee at resolution |
| Cost as share of capital at risk | High, and rises with leverage because notional rises faster than collateral | Fixed by the stake, because the stake is the capital at risk |
The asymmetry in the last row is worth sitting with. In a prediction market, spending $500 on a position means $500 is your maximum loss and the fees are a small fraction of that. In Perps, posting $500 of collateral at 20x controls $10,000 of notional — and every fee and every hour of funding is charged on the $10,000, not the $500. Leverage multiplies the cost base at the same time as it shrinks the buffer.
Detail on the Perps side is on fees and funding; the prediction market side is on Polymarket fees explained.
What they genuinely share
| Shared | Not shared |
|---|---|
| pUSD as the collateral asset | Account structure — Perps accounts key on your EOA base address, not a Safe address |
| Polygon for deposits and withdrawals | Risk model — leverage and liquidation versus capped loss |
| The Polymarket brand and account relationship | Cost structure — per-fill fee plus hourly funding versus a per-trade fee |
| Order-book matching | Settlement — continuous P&L versus $1/$0 at resolution |
Do not assume a balance is freely interchangeable between the two just because both use pUSD. The Perps account is funded through on-chain collateral deposits and is keyed to the EOA that signs the proxy creation, not to a Safe wallet address — even if you use a Safe elsewhere on Polymarket. Verify in the interface before planning around a shared balance.
Frequently asked questions
Does Polymarket prediction markets have leverage?
No. The prediction markets have no leverage and no margin call, and a position cannot be liquidated. Maximum loss is the price paid for the shares. Polymarket Perps is a separate product that does offer leverage — if you have read a "no leverage, no liquidation" statement about Polymarket and it did not name the product, it was describing the prediction markets.
Which product has better odds?
Neither — they price different things. A prediction market price is a probability. A Perps price is a market price of an underlying. "Better odds" is not a meaningful comparison between them; the meaningful question is which one can express your specific view without introducing failure modes your thesis does not depend on.
Is Perps riskier than the prediction markets?
By construction, yes, in three specific ways: leverage amplifies the effect of a given price move on your collateral, funding charges you for time, and forced liquidation can close a position that would have become profitable. That is not an argument against Perps — it is an argument for sizing and for choosing the product that matches the thesis.
Can I trade both from the same account?
Both are Polymarket products and both use pUSD. Account mechanics differ, and the Perps account is identified by your EOA base address. Rather than assume, check what the interface shows for each product and confirm transfers before relying on a shared balance.
Which gives me a bigger possible return?
Perps has no cap on percentage gain per position, because a leveraged directional move has no upper bound. A prediction market share caps at $1.00 — the maximum return is set by the entry price. But an uncapped return is not the same as a favourable one: the uncapped upside comes with a liquidation threshold that removes your capital outright. A 10¢ prediction share gives 10x with a capped loss; reaching the same multiple in Perps requires leverage that puts liquidation inside ordinary noise.
Where can I read the Perps mechanics in full?
Polymarket publishes Perps documentation at docs.polymarket.com/perps. On this site, start with the Perps hub, then liquidation, fees and funding, and leverage and sizing.
Curious about the venue? Polymarket lists both products. For the unleveraged path, the beginner's guide is the right starting point.