Prediction performance
SN21 models predict what each Google Ads change will do before the outcome is known. When it settles, every prediction is scored — this shows how reliable they are, one row per kind of change.
How to read this page
- What we measure
- The change's effect on cost, conversions and the goal metric — CPA (cost per lead) for lead-generation accounts, ROAS (return on ad spend) for ecommerce.
- At what level
- The campaign the change belongs to — not the whole account. An ad, keyword or ad-group change is measured on its campaign's results against a 60-day baseline.
- Over what window
- The next 7, 14, or 28 days after the change — switch the window on the table below.
Changes settled & scored
17,248
served changes settled since 20 August 2026
Kinds of change
14
one line each, below
Models compared
170
every model predicts every change
Best model today
UID 72
leaderboard #1
Where the number comes from
63 days, 2026-08-20 to 2026-10-21- 1. Captured1,127,131raw changes per day, before any rule
- 2. Measurable episodes52,144since 20 August 2026, after the published rules and 72-hour clustering
- 3. Served to models17,670of 23,587 eligible; 4,790 held out for validation, 200 on our own accounts
- 4. Settled & scored17,580this page — served changes old enough to have settled
- 5. Predictions scored4,264,976every model predicts every change
Each stage is deliberate. The page counts the fourth: a change reaches it about 15 days after it happened, once its first outcome window has closed and settled.
14 groups covering 93 change types. Open a row to see the types inside it.
Budget change — when a daily budget was raised or lowered 4,527 Reliable UID 91 6 types ›
Across 4,527 such changes (449,907 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Bid strategy switch — when the bid strategy was switched from one type to another 113 Reliable UID 165 8 types ›
Across 113 such changes (10,799 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Bid target change — when a bid-strategy target value changed (e.g. Target CPA $60 → $50) 488 Roughly right UID 173 13 types ›
Across 488 such changes (46,871 model predictions), a prediction here is roughly right. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Negative keyword changes — when negative keywords were added 2,694 Reliable UID 29 16 types ›
Across 2,694 such changes (262,012 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Keyword changes — when this changed 828 Reliable — 2 types ›
Across 828 such changes (80,762 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Other criterion changes (kind not recorded) — when this changed 1,733 Reliable UID 18 6 types ›
Across 1,733 such changes (164,694 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Location, audience and demographic targeting — when this changed 156 Reliable UID 12 3 types ›
Across 156 such changes (15,411 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Device, schedule and placement targeting — when this changed 162 Reliable UID 173 2 types ›
Across 162 such changes (14,180 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Roughly right Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Campaign pause / enable — when this changed 878 Reliable UID 72 1 type ›
Across 878 such changes (90,769 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Roughly right Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Ad group pause / enable — when this changed 114 Reliable UID 141 2 types ›
Across 114 such changes (10,780 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Other ad group changes — when this changed 168 Reliable — 3 types ›
Across 168 such changes (15,486 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Ad changes — when this changed 218 Reliable UID 38 5 types ›
Across 218 such changes (20,761 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Asset changes — when this changed 706 Reliable — 1 type ›
Across 706 such changes (58,724 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Combined changes — when several kinds of change happened together 4,463 Reliable UID 18 25 types ›
Across 4,463 such changes (434,281 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
14 groups covering 90 change types. Open a row to see the types inside it.
Budget change — when a daily budget was raised or lowered 4,169 Roughly right UID 91 6 types ›
Across 4,169 such changes (431,581 model predictions), a prediction here is roughly right. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Bid strategy switch — when the bid strategy was switched from one type to another 100 Reliable UID 126 7 types ›
Across 100 such changes (10,266 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Bid target change — when a bid-strategy target value changed (e.g. Target CPA $60 → $50) 444 Roughly right UID 165 13 types ›
Across 444 such changes (45,790 model predictions), a prediction here is roughly right. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Negative keyword changes — when negative keywords were added 2,355 Roughly right — 16 types ›
Across 2,355 such changes (242,842 model predictions), a prediction here is roughly right. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Keyword changes — when this changed 665 Reliable — 2 types ›
Across 665 such changes (68,504 model predictions), a prediction here is reliable. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Other criterion changes (kind not recorded) — when this changed 1,463 Reliable UID 126 6 types ›
Across 1,463 such changes (150,708 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Location, audience and demographic targeting — when this changed 114 Reliable UID 91 3 types ›
Across 114 such changes (11,798 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Device, schedule and placement targeting — when this changed 109 Reliable — 1 type ›
Across 109 such changes (11,202 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Campaign pause / enable — when this changed 877 Reliable — 1 type ›
Across 877 such changes (91,812 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Ad group pause / enable — when this changed 91 Roughly right UID 12 2 types ›
Across 91 such changes (9,326 model predictions), a prediction here is roughly right. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Other ad group changes — when this changed 144 Reliable — 3 types ›
Across 144 such changes (14,739 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Ad changes — when this changed 187 Reliable UID 55 5 types ›
Across 187 such changes (19,207 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Asset changes — when this changed 501 Reliable UID 126 1 type ›
Across 501 such changes (51,251 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Combined changes — when several kinds of change happened together 3,996 Reliable UID 246 24 types ›
Across 3,996 such changes (411,139 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
14 groups covering 88 change types. Open a row to see the types inside it.
Budget change — when a daily budget was raised or lowered 2,764 Roughly right UID 95 6 types ›
Across 2,764 such changes (284,990 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Bid strategy switch — when the bid strategy was switched from one type to another 66 Reliable UID 95 6 types ›
Across 66 such changes (6,716 model predictions), a prediction here is reliable. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Bid target change — when a bid-strategy target value changed (e.g. Target CPA $60 → $50) 332 Roughly right — 13 types ›
Across 332 such changes (34,141 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Negative keyword changes — when negative keywords were added 1,597 Roughly right UID 95 16 types ›
Across 1,597 such changes (163,702 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Keyword changes — when this changed 459 Roughly right — 2 types ›
Across 459 such changes (47,264 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Other criterion changes (kind not recorded) — when this changed 981 Roughly right UID 95 6 types ›
Across 981 such changes (100,815 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Location, audience and demographic targeting — when this changed 68 Roughly right UID 55 3 types ›
Across 68 such changes (6,950 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Roughly right.
Broken down by kind:
Device, schedule and placement targeting — when this changed 56 Reliable — 1 type ›
Across 56 such changes (5,661 model predictions), a prediction here is reliable. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Roughly right Range — how often the real result fell inside the predicted range
- Roughly right Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Campaign pause / enable — when this changed 476 Roughly right — 1 type ›
Across 476 such changes (48,999 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Ad group pause / enable — when this changed 73 Reliable UID 12 2 types ›
Across 73 such changes (7,473 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Other ad group changes — when this changed 100 Reliable — 3 types ›
Across 100 such changes (10,223 model predictions), a prediction here is reliable. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Ad changes — when this changed 134 Reliable UID 55 5 types ›
Across 134 such changes (13,711 model predictions), a prediction here is reliable. What that's made of:
- Reliable Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
Asset changes — when this changed 289 Roughly right — 1 type ›
Across 289 such changes (29,285 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Roughly right.
Combined changes — when several kinds of change happened together 2,539 Roughly right UID 95 23 types ›
Across 2,539 such changes (259,444 model predictions), a prediction here is roughly right. What that's made of:
- Roughly right Direction — called the change up or down correctly on the goal metric
- Reliable Size — how close the predicted size of the effect was
- Not reliable Range — how often the real result fell inside the predicted range
- Not reliable Goal — how close on the goal number itself (CPA or ROAS)
Today's leaderboard #1 (UID 72) on this change type: Reliable.
Broken down by kind:
A row reads Too early until at least 500 of its changes have settled. Open any row for the score breakdown and the best model.
Counted: changes to things that already exist, once their outcome has settled.
Not counted: brand-new campaigns & ads (no "before" to score), and recent changes still settling.
The winning model over time
The same winning model, week by week — is the best model on the subnet getting better? Every point is the mean of per-entry scores already published in the daily receipts, so anyone can recompute it.
- 7-day horizon
- 14-day horizon
- 28-day horizon
- week still open
| Week | Week ending | Horizon | Mean score | Entries | Status |
|---|---|---|---|---|---|
| 2026-W34 | 2026-08-23 | 7-day | 0.606 | 1086 | settled |
| 2026-W35 | 2026-08-30 | 7-day | 0.605 | 390 | settled |
| 2026-W35 | 2026-08-30 | 14-day | 0.592 | 1174 | settled |
| 2026-W36 | 2026-09-06 | 7-day | 0.591 | 861 | settled |
| 2026-W36 | 2026-09-06 | 14-day | 0.594 | 863 | settled |
| 2026-W36 | 2026-09-06 | 28-day | 0.518 | 489 | settled |
| 2026-W37 | 2026-09-13 | 7-day | 0.600 | 2502 | settled |
| 2026-W37 | 2026-09-13 | 14-day | 0.567 | 884 | settled |
| 2026-W37 | 2026-09-13 | 28-day | 0.580 | 1581 | settled |
| 2026-W38 | 2026-09-20 | 7-day | 0.601 | 2505 | settled |
| 2026-W38 | 2026-09-20 | 14-day | 0.577 | 2541 | settled |
| 2026-W38 | 2026-09-20 | 28-day | 0.585 | 874 | settled |
| 2026-W39 | 2026-09-27 | 7-day | 0.602 | 2467 | settled |
| 2026-W39 | 2026-09-27 | 14-day | 0.581 | 2536 | settled |
| 2026-W39 | 2026-09-27 | 28-day | 0.555 | 892 | settled |
| 2026-W40 | 2026-10-04 | 7-day | 0.605 | 2181 | settled |
| 2026-W40 | 2026-10-04 | 14-day | 0.582 | 2511 | settled |
| 2026-W40 | 2026-10-04 | 28-day | 0.563 | 2555 | settled |
| 2026-W41 | 2026-10-11 | 7-day | 0.635 | 2820 | settled |
| 2026-W41 | 2026-10-11 | 14-day | 0.590 | 2236 | settled |
| 2026-W41 | 2026-10-11 | 28-day | 0.575 | 2571 | settled |
| 2026-W42 | 2026-10-18 | 7-day | 0.617 | 3063 | settled |
| 2026-W42 | 2026-10-18 | 14-day | 0.602 | 2849 | settled |
| 2026-W42 | 2026-10-18 | 28-day | 0.579 | 2526 | settled |
| 2026-W43 | 2026-10-25 | 7-day | 0.612 | 750 | week still open |
| 2026-W43 | 2026-10-25 | 14-day | 0.595 | 1555 | week still open |
| 2026-W43 | 2026-10-25 | 28-day | 0.569 | 1286 | week still open |
Where these numbers come from
Every score on this page is the settle-flow score in the daily receipts, published per prediction at /v1/daily — the same numbers that feed the leaderboard. This page and the leaderboard can never disagree, because there is nothing here to disagree with: it is the same data, added up by change type.