# Polymarket 'By Date' Favourites at 84¢ a Month Out Came True 64% of the Time
Every settled Polymarket market worded 'by' or 'before' a date, January 2025 to April 2026, scored against its price 90, 30 and 7 days before the deadline and clustered by event. Markets above 65¢ a month out came true 64.3% of the time; a week out, 90.5%.
Canonical URL: https://vultax.com/research/polymarket-deadline-markets-accuracy
Author: [Vultax Research](https://vultax.com/editorial-policy)
Published: 2026-10-05
Last public revision: 2026-10-05
Published research snapshots and calculations; these exports do not contain current quotes or live account data. Observation windows and populations are stated in the source captions and methodology. Missing metadata is unknown, not zero. Figures describe Polymarket markets that settled between 1 January 2025 and 28 April 2026; a market price is a traded probability, not a forecast by Vultax, and nothing here is advice to trade.
## Key figures
- Deadline markets above 65¢ a month out that came true: 64.3% — 84 markets in 72 events, average price 83.7¢; 95% interval clustered by event 53.7–75.0%
- The same price band a week before the deadline: 90.5% — 74 markets in 70 events at an average 84.9¢
- Long shots at 5–15¢ a month out that came true: 13.7% — 190 markets at an average 8.9¢
- End-2025 markets open on 1 October 2025 that resolved Yes: 53 of 372 — Their 1 October prices added up to 70.5 expected Yes
## The short answer
A Polymarket market that asks whether something happens by a date comes true less often than its price says when it trades as a favourite a month before the deadline. Across 1,660 settled deadline markets with at least $50,000 traded, those priced above 65¢ thirty days out averaged 83.7¢ and came true 64.3% of the time, with a 95% interval of 53.7–75.0% when whole events are resampled.
Ninety days out the same band averaged 84.4¢ and came true 75.4% of the time. Seven days out it averaged 84.9¢ and came true 90.5%, above its price. A month out is where the price of 'it will happen in time' runs furthest ahead of the calendar.
This study scores only contracts with a deadline; [Vultax's calibration study](https://vultax.com/research/prediction-market-fair-value-calibration) tests Polymarket prices across resolved markets of every kind.
## Price against outcome, 30 days before the deadline
Each market is read at its last traded price nearest 30 days before its scheduled end, if it traded within two days of that point. 938 markets in 747 events qualify; together they averaged 20.9¢ and 18.2% resolved Yes.
Long shots under 5¢ almost never came true: one of 369. The 5–15¢ band came true more often than its price, and the expensive end is where the market overshoots.
Intervals resample whole events (2,000 draws), so rungs of one ladder move together.
| Price 30 days out | Markets | Events | Came true | 95% interval (by event) |
| --- | --- | --- | --- | --- |
| <5¢ | 369 | 332 | 0.3% | 0.0–0.8% |
| 5–15¢ | 190 | 171 | 13.7% | 8.7–19.1% |
| 15–35¢ | 171 | 151 | 21.6% | 15.3–28.6% |
| 35–65¢ | 124 | 107 | 42.7% | 34.1–51.8% |
| >65¢ | 84 | 72 | 64.3% | 53.7–75.0% |
[CSV](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.csv?dataset=by-price-table) · [JSON](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.json?dataset=by-price-table)
### Average price and share that came true, by price 30 days before the deadline
Polymarket deadline markets with $50,000+ volume, settled January 2025 to April 2026. Percent.
| Category | Average price 30 days out (%) | Share that came true (%) |
| --- | --- | --- |
| <5¢ | 2.2 | 0.3 |
| 5–15¢ | 8.9 | 13.7 |
| 15–35¢ | 24 | 21.6 |
| 35–65¢ | 48.4 | 42.7 |
| >65¢ | 83.7 | 64.3 |
[CSV](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.csv?dataset=by-price-chart) · [JSON](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.json?dataset=by-price-chart)
## Ninety, thirty and seven days out
Read at three distances from the deadline, the favourite band is priced about the same each time, but the share that came true drops at 30 days and ends above the price at 7. A market drops out of a horizon once it has settled, so each horizon reads only what was still open.
Days are counted back from each market's scheduled end. Above 65¢: markets (events) priced above 65¢ at that point, with the event-clustered interval for the share that came true. All: every market still open.
| Days out | Above 65¢ | 95% interval | All: price | All: came true |
| --- | --- | --- | --- | --- |
| 90 days | 57 (44 events) | 63.2–86.4% | 23.5¢ | 21.2% |
| 30 days | 84 (72 events) | 53.7–75.0% | 20.9¢ | 18.2% |
| 7 days | 74 (70 events) | 83.3–97.1% | 12.9¢ | 13.6% |
[CSV](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.csv?dataset=by-horizon-table) · [JSON](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.json?dataset=by-horizon-table)
### Markets above 65¢: average price and share that came true, by days to the deadline
Polymarket deadline markets with $50,000+ volume, settled January 2025 to April 2026. Percent.
| Category | Average price (%) | Share that came true (%) |
| --- | --- | --- |
| 90 days out | 84.4 | 75.4 |
| 30 days out | 83.7 | 64.3 |
| 7 days out | 84.9 | 90.5 |
[CSV](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.csv?dataset=by-horizon-chart) · [JSON](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.json?dataset=by-horizon-chart)
## Why the count is by event, not by market
A deadline ladder lists one question at several dates, and its rungs settle together: if the first rung comes true, every later one does too. The 1,660 markets belong to 1,141 events, so every interval on this page resamples whole events rather than single markets.
Weighting each event once instead of each market moves the 30-day favourite figure to 67.8%. The largest single event in that band holds three markets, so no one ladder drives the result.
## The end-2025 settle, scored from 1 October 2025
1,052 markets with $50,000 or more traded were scheduled to end on 31 December 2025 or 1 January 2026. 372 of them, in 273 events, were still trading on 1 October 2025, the same calendar point as the end-2026 reading below.
At their 1 October prices they averaged 18.9¢. 53 resolved Yes (14.2%, interval 10.0–19.1%). Those above 65¢ on 1 October came true 74.2% of the time at an average of 85.1¢.
| Price on 1 Oct 2025 | Markets | Average price | Came true |
| --- | --- | --- | --- |
| <5¢ | 105 | 2.8¢ | 0.0% |
| 5–15¢ | 142 | 8.2¢ | 4.2% |
| 15–35¢ | 62 | 22.8¢ | 22.6% |
| 35–65¢ | 32 | 47.9¢ | 31.2% |
| >65¢ | 31 | 85.1¢ | 74.2% |
[CSV](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.csv?dataset=year-end-2025-table) · [JSON](https://vultax.com/research/polymarket-deadline-markets-accuracy/data.json?dataset=year-end-2025-table)
## The end-2026 markets, fixed before they settle
Vultax's ledger records every open Polymarket market scheduled to end on 31 December 2026 or 1 January 2027 in one run at 00:00 UTC on 1 October 2026, the same calendar point as the 2025 reading above. That run fixes the cohort: its SHA-256 is added to this page as a dated update, and the markets are scored with the same price bands and event-level intervals in the first week of January 2027.
## Method and limits
The universe is every market in the Polymarket-v1 fill tape (settled markets, 1 January 2025 to 28 April 2026) whose slug says 'by' or 'before' a date or 'in' a year, with price-target ladders on crypto and assets excluded and at least $50,000 traded. The tape's settled result matched Polymarket's final prices in 2,326 markets, with no disagreements.
Horizons count back from each market's scheduled end date. A market whose event happened early settles early and leaves the later horizons, which is why the share of all markets that came true is lower at 7 days than at 90. Intervals are 95% percentile intervals from 2,000 event-level bootstrap draws.
## Questions
### What is a Polymarket deadline market?
A market that asks whether something happens by or before a date, such as a ceasefire by 31 December. Many are listed as ladders: one question at several dates, each settling Yes once the event happens in time.
### Are Polymarket 'by date' odds accurate?
At the extremes and close to the deadline, yes. A month before the deadline the favourite band overshoots: markets above 65¢ averaged 83.7¢ and came true 64.3% of the time.
### How often does a Polymarket long shot on a deadline come true?
Under 5¢ a month out, almost never: one of 369. Between 5¢ and 15¢, 13.7%.
### Why are deadline markets counted by event rather than by market?
The rungs of one deadline ladder resolve together, so counting them as independent markets would overstate how much evidence there is.
### How will the Polymarket markets ending in 2026 be scored?
Every open market scheduled to end on 31 December 2026 or 1 January 2027 is recorded at 00:00 UTC on 1 October 2026 and scored in the first week of January 2027 with the same price bands and intervals.
## Sources and methodology
- [Boka Qin and Rui Yang — Polymarket-v1 Database (TimeSeventeen)](https://huggingface.co/datasets/TimeSeventeen/Polymarket-v1) — Settled-market fills and aligned market metadata, 1 January 2025 to 28 April 2026, published under CC BY 4.0. Vultax built the hourly price panel and the statistics.
- [Polymarket Gamma API — markets and events](https://docs.polymarket.com/) — Event membership (for clustering), settled prices and the end-2026 snapshot.