Every alerted position, graded
Each row is a position the alert feed flagged, with its price at the alert and its grade once the market resolved. Losses stay on the page.
Both intervals account for positions that share a market, since those settle together.
- 617 won
- 379 lost
- 21 voided
- 105 sealed until resolution
Sealed rows count in the total from the moment they are flagged, and move to won, lost or voided when their market settles.
Graded positions tagged Culture
Newest first. Category tags open a filtered view.Showing the 2 graded rows tagged Culture. The counts above cover the whole ledger. View the full ledger
| Detected | Market | Position | At alert | 5+ min | +1h | +24h | Outcome |
|---|---|---|---|---|---|---|---|
| 2026-07-04 | Will Gigi Hadid be one of Taylor Swift's bridesmaids?Culture | Bought No$100 | 99¢ | 98¢ (-1) | 99¢ (0) | 100¢ (+1) | won2026-07-04 |
| 2026-07-04 | Will Selena Gomez be one of Taylor Swift's bridesmaids?Culture | Bought No$400 | 71¢ | 95¢ (+24) | 98¢ (+27) | 100¢ (+29) | won2026-07-04 |
Get the next conviction alert, typically within minutes
Pro sends conviction alerts on Telegram and email when a tracked forecaster makes a move that clears the conviction gate, typically within minutes. Each wallet and each category has a daily limit, and alerts held back by these limits go into the daily digest. Pro also opens every skill score on the full leaderboard. The receipts ledger stays free.
How the ledger is kept
Frozen at the first flag
Each row keeps the date and price of the position’s first alert. A wallet adding enough to re-flag it updates the row’s size and flag count (published at resolution), and no new row is added.
Graded by the market
Every row settles mechanically on the market’s own resolution, and nothing is edited after the fact. A voided row stays on the ledger and is left out of the win rate and the price comparisons.
Measured against the price
A win rate says little until you know what the prices implied. The cards above compare the share won with the win chance implied at each alert, and at the first trade five or more minutes later, a delay meant to leave time to read an alert and place an order. That trade can be a sale or a purchase. At the taker price (the first purchase at the asking price five or more minutes after each alert), the 878 alerted positions graded by 2026-09-28 won 1.74 percentage points more often than that price implied, before fees, with a 95% interval from −0.96 to +4.44 that includes zero (a pre-registered run, on the research page).
Data notes
- Prices
- The At alert, +1h and +24h prices are the last trade at or before each horizon. The 5+ min price is the first trade at least five minutes after the alert, and in a thin market that trade can come hours later. All come from Polymarket’s public fills, before fees.
- The 95% intervals
- Differences are in percentage points (pts). An interval that includes zero means a difference that size is within what chance produces at those prices, and one entirely above zero means a larger difference than chance usually produces. Positions on the same market settle on one result, and the intervals count them that way. Two alerts on the same side of a market count like one bigger bet, and a pair on opposite sides adds no uncertainty, because exactly one of them wins. Separate markets count as independent. Markets tied to one event, such as a match and its single games, can settle in related ways, and the intervals do not allow for that.
- Sealed rows and the ledger identity
- Unresolved alerts are sealed. Each one is counted the moment it is flagged, and its full row publishes when the market resolves, with the original detection date frozen. Their details stay sealed until the market resolves because realtime alerts are the paid product. 1,122 alerted positions = 996 resolved (617 won, 379 lost) + 21 voided + 105 sealed.
- Both sides alerted
- A market alerted on both sides before the change on September 24, 2026 carries a row for each side, each graded on its own and tagged “both sides alerted”. Since the change, OVERROUND does not alert both sides of one market (the methodology gives the full rule).
- Dates, sizes and sides
- Dates carry day precision and position sizes are rounded. The date beside a grade is the market’s scheduled close date, shown once it has passed. Grades settle on the market’s recorded outcome, independent of that date. “Bought” and “Sold” describe the tracked wallet’s own action at detection.
- Ledger and leaderboard
- A category can appear in this ledger without appearing on the leaderboard. Alerts fire on live wallet moves, while the board lists only wallets whose records clear its public evidence floor of resolved events.
- Method
- Alert timing, gating, and grading are documented in the methodology. Validation work is pre-registered on the research page, and aggregate calibration and grading statistics are on the statistics page. Any wallet’s own record can be tested against its entry prices with the free wallet checker.
The Weekly Skill Report. Each Monday, one free email counts the positions the alert feed first flagged last week and how many have since won or lost. It names the busiest categories and the ledger’s running total, and links to every graded row.
Free. Unsubscribe anytime. Privacy.