Why Polymarket’s Closing Price Can Differ From Your Chart

Polymarket’s five-minute crypto Up/Down markets now use a 60-second Chainlink time-weighted average price—not the 30-second window described in the initial August announcement. Polymarket introduced TWAP settlement on August 7, 2026, then lengthened the five-minute markets’ window on August 14 at 00:00 UTC. The published rules as of October 5, 2026, make the designated average-price comparison—not the last price on your preferred Bitcoin chart—the settlement test. Source: Polymarket’s changelog.

The calculator below illustrates the distinction using the draft’s hypothetical price path. It does not reproduce Chainlink’s production feed or predict an actual market’s resolution.

Change the opening reference and two prices; expand the timing control to test how a late move affects the average.

Five-Minute Markets Changed Twice In August

TWAP means time-weighted average price: a price averaged across a lookback period, rather than a single instantaneous observation. A price that persists for more of that period receives more weight than one that appears only briefly.

The August 7 announcement and the subsequent five-minute update describe different windows. Reading only the initial announcement leaves you with an outdated rule for the five-minute contracts.

Crypto Up/Down Duration	August 7 Rollout	Window As Of October 5
Five minutes	30 seconds	60 seconds
15 minutes	60 seconds	60 seconds
Four hours	60 seconds	60 seconds

Polymarket says both the opening “price to beat” and the final settlement price come from the applicable TWAP feed. Its current documentation describes a fixed 60-second crypto TWAP lookback. That is not an average over the entire five-minute, 15-minute or four-hour contract. Sources: rollout and subsequent update, current TWAP documentation.

Contract duration and averaging-window length are separate settings. “Five minutes” identifies the market’s duration; “60 seconds” identifies the reference price’s lookback. The longer contracts listed above do not use an average spanning their full duration merely because they last longer.

This change concerns the specified crypto Up/Down markets. It does not establish the settlement method for every Polymarket crypto question, much less an election or sports contract. For an individual trade, the market’s own rules must identify the source and comparison you are buying exposure to.

The Opening Reference Is Also An Average

The change is not simply “use an average at the end.” The opening price to beat is also taken from the applicable TWAP feed. Comparing a closing TWAP with an opening spot-chart candle therefore substitutes a different starting reference for the one the rules designate.

A September 20, 2026 Bitcoin five-minute market explicitly names the Chainlink BTC/USD 60-second TWAP stream as its resolution source. Its rules assign “Up” when the relevant closing TWAP is greater than or equal to the opening reference; otherwise, “Down.” They also warn that other sources and spot markets do not determine resolution. Source: the market’s rules.

That wording establishes two separate checks for this example market: which prices count, and how those prices are compared. The specified stream supplies the reference; the greater-than-or-equal condition assigns the outcome. Neither part can be replaced by the appearance of a chart candle.

The equality rule matters when the two references match exactly. Under the quoted rule, equality resolves to Up, not Down. Do not assume that a different market has the same tie treatment without reading its wording.

For a beginner, the practical task is to identify the actual opening reference before interpreting a late move. A chart can show the asset rising while the designated closing average remains below the designated opening average. That is a disagreement between measurements, not necessarily an inconsistency in settlement.

A Late Spot Rally Can Still Resolve Down

Consider one hypothetical market with an opening TWAP—the price to beat—of $100,000. During the final 60-second lookback, assume the underlying price being averaged stays at $99,990 for 55 seconds, then $100,050 for five seconds.

This simplified illustration uses two constant prices. It is not a reconstruction of Chainlink’s production calculation, sampling or feed implementation.

The closing time-weighted average is ($99,990 × 55 + $100,050 × 5) ÷ 60 = $99,995.

Measurement	Value	Against Opening Reference
Opening TWAP	$100,000	Price to beat
Last price	$100,050	Above
Closing TWAP	$99,995	Below

The last price is above $100,000, but the average is below it. Under the assumed rules, the result is Down. If the closing TWAP exactly equalled the opening reference, the quoted rule would instead produce Up.

The late move does affect the average. It simply does not replace the price that occupied the earlier portion of the lookback. In this example, the higher price contributes for five seconds, while the lower price contributes for 55 seconds. Treating the last observation as the settlement price discards that difference in duration.

The interactive example lets you change the two prices and the time spent at the later price. Its output answers only a conditional question: what would this simplified average imply under the example market’s comparison rule? It cannot confirm the actual Chainlink reference at a particular market’s close.

A chart screenshot showing a final price above the opening reference is therefore insufficient to establish an Up result. You would need the designated closing TWAP and the correct opening reference. Equally, an illustrative calculation is not evidence that a live market settled incorrectly; its assumed price path may not match the designated feed.

Spot Price, Settlement Reference And Share Price Are Different

Keep three numbers separate: the underlying asset’s spot price, the settlement-reference TWAP, and the price of an Up or Down share.

The spot price describes the asset at a particular source and moment. It can help you follow a move, but the example market’s rules expressly say that other sources and spot markets do not determine resolution.

The settlement-reference TWAP is the price measurement named in the contract. The opening and closing references determine which side satisfies the comparison rule. This is the number to investigate when a spot chart and the expected outcome disagree.

The share price is what you can pay or receive for a position. An Up share trading at 60¢ is not a Bitcoin price and is not the closing TWAP. It is a transaction price for the contract’s payoff. It can change while the underlying reference remains on the same side of the opening threshold.

This distinction also separates two decisions. The resolution question is whether the designated reference meets the rule. The trading question is whether the available share price and costs justify taking that exposure. Correctly understanding the first does not automatically make the second attractive.

Averaging Reduces A Snapshot’s Sensitivity, Not All Risk

CoinDesk’s August 7 report connected the change to trader complaints and research identifying activity consistent with settlement manipulation. The concern was structural: a short-lived move in the underlying asset could change a binary payout when settlement relied on a price snapshot. Source: CoinDesk.

The June 30, 2026 research preprint, Settlement Manipulation in Prediction Markets, found settlement-time spot order-flow spikes and subsequent reversals after Polymarket’s five-minute Bitcoin contracts launched. It found manipulation largely absent in the 15-minute contracts studied. Source: original research.

That comparison concerns contract duration, not the later TWAP-window change. It does not establish that all longer-duration markets are safe from manipulation, or measure the effectiveness of the subsequently introduced 60-second settlement window.

The evidence also did not directly establish that the spot traders and prediction-market holders were the same people or prove intent, as CoinDesk notes. Activity consistent with manipulation is not the same claim as identifying a particular actor’s positions and purpose.

The arithmetic explains why averaging changes the vulnerability: a brief move contributes only for the portion of the lookback during which it occurs. But averaging does not establish that manipulation is impossible. When the average is close to the threshold, even a limited contribution can matter to the binary outcome.

The sources retained here do not supply a measured reduction in manipulation attributable to the 60-second window. Treating the change as a complete fix would go beyond that evidence. The supported claim is narrower: Polymarket changed the settlement reference, and the longer average gives a brief move less weight than a single snapshot would.

The Ask And Fees Determine Your Outlay

Continue the hypothetical example. Suppose Up shares have a 58¢ bid and 60¢ ask, with enough shares quoted to fill your entire 10-share order. The bid is the available selling price; the ask is the available buying price.

Buying 10 shares at the ask costs $6, not the 59¢ midpoint multiplied by 10. The midpoint lies between the quoted prices, but this example does not offer execution there. Using it as your purchase price understates the assumed cost.

Assume you buy as a taker, using the crypto fee coefficient of 0.07 published as of October 5, 2026, with no rebates or other costs. Polymarket’s fee formula gives 10 × 0.07 × 0.60 × (1 − 0.60) = $0.168. Source: fee formula.

Item	Amount	Assumption
Share purchase	$6	10 shares at 60¢
Taker fee	$0.168	Coefficient 0.07
Total outlay	$6.168	No other costs
Winning payout	$10	$1 per share

If Up wins, the $1-per-share payout totals $10, leaving $3.832 net after the purchase and assumed fee. If Down wins—as in the illustrative price path above—you lose the $6.168 outlay. Source: share payout.

Selling immediately at the unchanged 58¢ bid would instead lose 20¢ before fees. That is the spread cost in this hypothetical round trip; it is not the fee-inclusive loss and does not depend on waiting for settlement.

These figures assume the entire order fills at the stated price. They do not describe the cost of an order that exceeds the available quantity at that quote. The fee coefficient is also the dated assumption used here, not a reason to skip checking the fees applicable to your own trade.

Verify The Contract Before Buying The Outcome

For the resolution check, read the exact market’s source, averaging window, opening reference, end time and equality rule together. A remembered announcement can be stale; a familiar spot chart can use the wrong measurement; a correct closing reference can still be compared with the wrong opening value.

For the execution check, inspect the ask you can buy, the quantity available there and the applicable fees. The hypothetical 60¢ purchase works only because the stated quote can fill all 10 shares. A displayed probability or midpoint does not replace those execution details.

Use the broader pre-trade checklist if either check is unfamiliar. The TWAP change tells you what the specified contracts compare. It does not make a short-duration position low-risk, guarantee a favorable fill, or turn the final spot tick into the winning settlement reference.