Polymarket Payout Time 2026: 37,659 Timed
Polymarket payout time measured on 37,659 markets: the median wait past a deadline is 2h17m, sport takes 3h56m and crypto settles in 54 seconds.
Affiliate disclosure: contains affiliate links; we may earn a commission at no extra cost to you. 18+ only — see BeGambleAware.org. Last updated 21 October 2026.
Key takeaways
- The real Polymarket payout time is 2 hours 17 minutes at the median, measured on every market that settled at or after its own deadline. Not the 2 hours everyone quotes — that is the dispute window, and it is a floor, not an average.
- Only 45.8% of markets beat two hours, and just 27.6% of the money does. Weight by volume instead of by market count and the familiar figure describes barely a quarter of the exchange.
- There are two settlement rails and they are 250 times apart. Oracle-fed crypto markets resolve in a median 54 seconds. Everything on the human-proposal rail takes a median 3 hours 53 minutes.
- Sport is the slow giant. It is $2.1bn of the volume we timed and it settles at a median of 3 hours 56 minutes, with 66.8% of markets landing in a clump between three and five hours.
- Disputes are rare and expensive in time. 62 of 37,659 markets were disputed — 0.165% — but 21 of those took over a day and the worst ran 14.7 days.
On this page
Ask how long Polymarket payout time takes and every page you find says about two hours. That number is not a measurement. It is the length of the dispute window written into the oracle contract, and it has been copied from the documentation into a hundred guides without anybody checking what it means for the person refreshing a settled market.
It matters because the wait is where the money sits. A market that ended an hour ago is holding your stake, and on the biggest category on the exchange you are not getting it back inside two hours — you are getting it back in roughly four.
So we timed the exchange. We pulled every market with at least $10,000 of volume scheduled to end in the 30 days to 22 October 2026 — 37,659 markets holding $3.72bn — and measured the gap between each one's stated deadline and the moment its result went final. Every figure below is that gap, read from Polymarket's own API, timed off the stamp the exchange writes when resolution completes.
What the Polymarket payout time actually is
29,520 of those markets settled at or after their own deadline. Here is the whole distribution, in the only unit that matters — how long you wait after the thing you bet on is over.
29,520 markets that settled at or after their stated deadline, from a 37,659-market sample ending 22 September to 22 October 2026, each with $10,000+ of volume.
Two things jump out. The median is 2 hours 17 minutes, which looks like a vindication of the folklore until you see the shape around it: the 25th percentile is 79 seconds, the 75th almost four hours. There is no typical market — only fast ones and slow ones.
The second is the gap between counting markets and counting money. 45.8% of markets settled inside two hours, but those markets held only 27.6% of the $2.32bn in the group. The fast half of the exchange is the cheap half. Weight by stake and the two-hour claim covers barely a quarter of it.
Two settlement rails, not one
The bimodal shape has a clean cause, and it is the most useful thing here: Polymarket settles on two different mechanisms, and 11,018 of our 37,659 markets never touched the oracle process the guides describe.
Those 11,018 have no proposal history in their resolution record. Every one of them is a crypto price market, and every one names a Chainlink data stream as its resolution source. They are settled from a price feed by machine. The other 26,641 go through UMA's optimistic oracle, where a human or a bot has to post a proposed answer and a bond before any clock starts.
Medians over markets settling at or after their deadline. The oracle rail is tight: 5th percentile 53 s, 95th 87 s, slowest of 11,018 markets 5 min 17 s.
That is a 250-fold difference in median wait between two products sold on the same site. If your experience of Polymarket is five-minute Bitcoin markets, settlement feels instant and the two-hour figure sounds pessimistic. If you trade football, it has never once been two hours.
How long Polymarket takes by market type
Each market carries a fee-schedule tag that doubles as a category label, so we can split the sample by what you are actually betting on.
Market counts cover the whole sample; wait percentiles use only the markets in each category that settled at or after their deadline. Categories under 150 markets are omitted.
Politics is the slowest thing with real money in it — a median of 16 hours and a 90th percentile of nearly a day and a half, on $183m. A political question needs an official result somebody will bond against, and those arrive on their own schedule. Weather is worse at 11 hours, but it is $24m, so almost nobody feels it.
The giveaway is the "within 2 hours" column. Every category on the UMA rail is in single digits except finance prices. The two-hour figure is not a description of any of them; it is a description of the crypto rail, generalised.
Why the two-hour Polymarket payout time is the wrong number
Polymarket's own resolution documentation describes a two-hour challenge period: once an outcome is proposed, anyone can dispute it inside that window, and if nobody does, the proposal becomes the result. That is accurate and it is where the number comes from. The error is treating a window that opens on the proposal as if it opens on the deadline.
Markets can override that window length, which lets us test which half of the wait is doing the work. If the challenge period were the bottleneck, a shorter window would mean a shorter wait. On the UMA rail it does not.
UMA-rail markets settling at or after their deadline, grouped by the market's own challenge-window setting.
Read the 15-minute row against the 60-minute row. Markets with a quarter-hour challenge window took a median 3 hours 21 minutes; markets with a window four times longer took 43 minutes. And 851 markets with the window set to zero — no challenge period at all — took a median of 8 hours 38 minutes.
The window is not the clock. The clock is how long it takes somebody to propose an answer, and shortening the challenge period on a market whose result nobody can confirm for four hours changes nothing. That is why the guides are wrong without being inaccurate: they quote a real parameter that does not govern the thing readers are asking about.
Sport: the four-hour clump
Sport is where most of the money and nearly all of the waiting lives — $2.3bn across 19,087 markets in our sample. Its distribution is the strangest in the dataset, because it is barely a distribution at all.
11,871 sport markets that settled at or after their deadline. 66.8% of them landed between three and five hours — the tightest band in the whole sample.
Eighty per cent of sport markets settle in a two-hour band between 2h17m and 4h22m. That is not what per-match resolution looks like; it is what a scheduled job looks like — proposals swept in batches rather than fired at the final whistle. It is also why sport's "within 2 hours" figure is 3.0% and not something nearer half.
The practical upshot is simple and nobody tells you it: if you backed a football result, plan on four hours, not two. Our largest sport market, $13.5m on a Norway fixture, took 3 hours 58 minutes. Size does not buy speed either — across the 502 markets over $500,000 in our sample the median wait was 3 hours 36 minutes, worse than the exchange as a whole.
The fifth of markets that pay before the deadline
The flip side is the part nobody mentions at all. 8,139 of our 37,659 markets — 21.6%, holding $1.40bn — resolved before their own stated end date, at a median of 68.6 hours early.
Share of each category's markets that went final before the stated end date. The largest case, a $24.5m Russian parliamentary market, settled 203 hours early.
This is the kindest fact about the exchange and it has a sharp edge. On a conditional question — will X happen by 31 December — the end date is a deadline for the question, not a commitment to hold your capital that long. If the answer arrives early, the market closes early and your shares redeem early.
It cuts the other way too. Because the end date is a placeholder on most sport markets, the countdown on the page is not a payout estimate — it is the latest the question could stay open.
What a dispute costs you in time
A dispute is the tail risk on every UMA-rail market: somebody bonds against the proposed answer, the proposal is voided and the clock restarts. A second disputed round escalates to a token-holder vote the documentation puts at roughly 48 hours. Here is how often that happens, and what it costs.
Disputes identified from each market's recorded resolution history. Undisputed markets on the same rail settled at a median of 3 h 52 min.
The median disputed market resolves in under three hours, which is the surprise. Most disputes are a bot correcting a bot, settled inside the next round. The damage is all in the tail: half of the disputed markets that went past a day went past two, and one ran for a fortnight.
The ugliest case was a $2.0m market on whether Bitcoin finished 5 October up or down — a question with an objectively checkable answer. It logged ten dispute entries and took 56 hours 30 minutes. When a dispute war starts, the clarity of the question is no protection.
One correction while we are here, because three sources give three answers. Polymarket's documentation says the proposal bond is "typically $750" and the best third-party guide says $500. In our sample the bond field reads 250 on 24,660 markets and 500 on 1,981, and the proposer reward is most often 0.8, not the $2 or $5 usually quoted. Bonds are set per market, and mostly smaller than the guides claim.
The markets that take days
273 markets in our sample took more than 24 hours, holding $39m between them. They are not random: 63 carried no category tag at all, and the rest are concentrated in culture, weather and politics. Here are the slowest with real money in them.
Slowest markets over $100,000 of volume, from markets scheduled to end in the 30 days to 22 October 2026.
Every one has the same shape: a question whose answer depends on a definition rather than a scoreboard. "Returns to normal", "effectively closed", "invokes war powers" — there is no feed to read and no body that publishes a yes. A proposer must make a judgement and bond it, and when the judgement is contestable nobody goes first.
$12.3m sat in the Hormuz market for five and a half days past its deadline. That is the real resolution risk on Polymarket, and it is not the oracle failing — it is a question written in words that do not resolve themselves.
Check the resolution source, not the countdown
Every market page names the source its result is read from. A price feed means seconds, a scoreboard means hours, and a sentence containing "effectively" or "returns to normal" means days. That line predicts your wait better than any figure here.
Visit Polymarket.com →Affiliate link. 18+ only. Polymarket restricts access from the United States, Canada and other listed jurisdictions. Prediction markets carry a substantial risk of loss. Nothing here is financial advice.
Resolved is not withdrawn
Every number above measures one thing: when the result went final and your winning shares became worth $1 each. That is not money in your bank, and the pages ranking for "Polymarket payout" mostly answer the second question while appearing to answer the first.
There are three clocks, and only the first one is what we measured:
One: resolution. The deadline passes, somebody proposes, the challenge window runs, the result goes final. Median 2 hours 17 minutes on markets that waited at all; 54 seconds on the crypto rail; days on a definitional question.
Two: redemption. Winning shares become redeemable for $1.00 and the balance lands in your Polymarket account. There is no expiry on this, so an unredeemed position is not a lost one.
Three: withdrawal. Getting the balance off the platform, which is a crypto-rail question about networks and fees, not an oracle question. We cover that separately in our Polymarket deposit and withdrawal costs guide.
Conflating one and three is why the published advice is so unhelpful. "Crypto withdrawals take minutes" is true and irrelevant to somebody whose football market ended four hours ago. If your balance has not moved, check whether the market has resolved at all before you go looking at network congestion.
My verdict on Polymarket payout speed
Polymarket settles faster than its reputation and slower than its documentation implies, and the two-hour number is the worst possible summary because it happens to sit near the median while describing almost nothing. Plan on 90 seconds for a crypto price market, four hours for anything with a scoreboard, a day for politics, and a week for any question written in adjectives.
By the standards of the mechanism, that is good. A sportsbook has a trading desk reading a feed; Polymarket has a bond and a dispute window, and the price of settling without a trusted operator is that somebody must volunteer to go first. 99.1% inside 24 hours is a strong result for a market with no referee.
What this cannot tell you is whether the wait is getting better or worse: one 30-day window measures a level, not a trend. And these are the markets that resolved, so a question still stuck in a dispute is absent by construction and the true tail is fatter than ours. We also filtered to $10,000 of volume and up, and the tiny markets we excluded are likely slower, not faster.
The one habit worth taking from all this: read the resolution source line on the market page before you read the odds. It is the single best predictor of how long your money will be locked after you are right, and it is the field nobody looks at.
Related guides
- How Polymarket Markets Resolve: UMA Disputes Explained — the mechanism behind every number above, including what happens when a dispute reaches a vote.
- Polymarket Fees 2026: Every Rate by Category — the fee schedules that give each market the category tag we grouped by.
- Polymarket Deposit Fees 2026: The Real Cost — the funding and withdrawal legs, which are the third clock in the section above.
- Polymarket Arbitrage 2026: 716 Fields Priced — why a locked dollar is worth less than it looks once settlement time is capital time.
- Polymarket Odds Accuracy 2026: 9,836 Scored — how good the prices are before any of this waiting starts.
- How Does Polymarket Work? Explained for 2026 — the share mechanics that make a resolved market worth exactly $1.
Polymarket payout time FAQ
How long does Polymarket take to pay out?
A median of 2 hours 17 minutes after the market's stated deadline, across 29,520 markets that settled at or after it in the 30 days to 22 October 2026. The spread is wide: 45.8% inside two hours, 93.7% inside six, 99.1% inside a day. Which end you land on depends on the market type.
Why do Polymarket crypto markets settle so much faster?
Because they skip the oracle entirely. 11,018 markets in our sample had no proposal in their resolution record, and all of them read a Chainlink price feed instead. They settled at a median of 54 seconds and a 95th percentile of 87 seconds. Everything else waits for a human or a bot to propose an answer.
Is the two-hour Polymarket resolution window a payout time?
No. Two hours is the challenge period, and it starts when somebody proposes an outcome, not when the market closes. The proposal is the bottleneck: markets with a 15-minute window took a median 3 hours 21 minutes, those with a 60-minute window 43 minutes.
How long do Polymarket sports markets take to settle?
A median of 3 hours 55 minutes across the 11,871 sport markets we timed, and only 2.9% settled within two hours. Sport is unusually consistent: 66.8% of them landed between three and five hours of their deadline, which looks like batched proposals rather than per-match settlement. Plan on four hours.
How often are Polymarket markets disputed, and how long does that take?
62 of 37,659 markets, or 0.165%, had a dispute recorded — 0.23% of the markets on the oracle rail. The median disputed market still resolved in under three hours, but the tail is brutal: 21 took over a day, 13 over two days, and the worst ran 14.7 days. Half of the disputed markets were disputed more than once.
Can a Polymarket market resolve before its end date?
Routinely. 8,139 of our 37,659 markets — 21.6%, holding $1.40bn — went final before their stated end date, a median of 68.6 hours early. Culture did it 47.6% of the time and sport 40.9%. On sport the end date is a generous placeholder, so the countdown is the latest the question can stay open, not your wait.
What is the slowest a Polymarket market can take to pay out?
The slowest in our sample was 14.1 days, on a question about whether Trump would invoke war powers in Yemen. The pattern is consistent: the slow markets are the ones whose criteria rest on a definition rather than a scoreboard. A $12.3m market on whether Strait of Hormuz traffic had "returned to normal" took five and a half days past its deadline.
Do I have to claim Polymarket winnings, or do they arrive automatically?
Winning shares become redeemable for $1.00 each once the result is final, with no expiry, so an unclaimed position is not a lost one. Withdrawal is a separate step governed by networks and fees, not the oracle — which is why "crypto withdrawals take minutes" explains nothing about an unpaid settled market.
Method and sources
We pulled every closed Polymarket market with at least $10,000 of volume whose stated end date fell between 22 September and 22 October 2026, day by day, from the Gamma API — 37,659 unique markets holding $3.72bn, read on 21 October 2026. Wait time is each market's recorded close timestamp minus its own stated end date. That close timestamp is the resolution timestamp, not a trading cut-off: in a separate unfiltered pull of 2,100 consecutively closed markets it was identical to the recorded oracle end time on all 2,100. Category splits use each market's live fee-schedule tag, challenge-window splits its own override field, and disputes are counted from its recorded resolution history. Percentiles cover markets that settled at or after their deadline, with the early population analysed separately, because mixing the two produces a median that describes neither. One 30-day window is a measurement, not a trend, and nothing here is financial advice.
- Polymarket docs: resolution — the two-hour challenge period, the proposal and dispute bonds described as "typically $750", and the escalation path to a token-holder vote.
- UMA docs: how the optimistic oracle works — the propose-and-challenge design, and the Data Verification Mechanism vote a twice-disputed question escalates to.
- Chainlink Data Streams — the price-feed product named as the resolution source on all 11,018 markets that settled without a proposal.
- Polymarket docs: fees — the per-category fee schedules whose tags we use as category labels throughout.
- Gamma API: /markets — the endpoint every timestamp, volume, bond and category figure above is read from.