Put the forecasts on the line.
Five models. Equal starting credits. The weather decides who gets paid. Everyone can watch; only the models participate.
UTC · common deadline for scheduled opponents
0 contests with provisional observations
0 voids · 0 missed execution windows
The model bankrolls
| MODEL | AVAILABLE | RESERVED | EQUITY AT COST | REALIZED P&L | OPEN | W–L |
|---|---|---|---|---|---|---|
| AIFS Single | 11,345.04 | 0.00 | 11,345.04 | +1,345.04 | 0 | 287–210 |
| ICON | 10,241.01 | 0.00 | 10,241.01 | +241.01 | 0 | 272–204 |
| WeatherNext 2 | 9,987.55 | 0.00 | 9,987.55 | -12.45 | 0 | 215–212 |
| IFS HRES | 9,518.56 | 0.00 | 9,518.56 | -481.44 | 0 | 191–313 |
| IFS Ensemble | 8,907.84 | 0.00 | 8,907.84 | -1,092.16 | 0 | 249–275 |
Equity at cost = available + reserved credits. There is no mark-to-market profit. Opponents and exposure limits affect trading outcomes.
Realized profit
Settlement values · latest 30
| FINALIZED UTC | AIFS | EPS | WN2 | HRES | ICON |
|---|---|---|---|---|---|
| Oct 11 · 05:06 UTC | +1,338.35 | -1,123.52 | -73.68 | -442.91 | +301.77 |
| Oct 11 · 05:06 UTC | +1,338.35 | -1,107.80 | -50.63 | -465.96 | +286.05 |
| Oct 11 · 05:06 UTC | +1,331.19 | -1,107.80 | -43.48 | -465.96 | +286.05 |
| Oct 11 · 05:06 UTC | +1,342.87 | -1,130.71 | -20.57 | -465.96 | +274.38 |
| Oct 11 · 05:06 UTC | +1,320.94 | -1,130.71 | -46.41 | -444.04 | +300.22 |
| Oct 11 · 05:06 UTC | +1,315.44 | -1,125.21 | -46.41 | -412.27 | +268.45 |
| Oct 11 · 06:09 UTC | +1,312.55 | -1,122.32 | -46.41 | -412.27 | +268.45 |
| Oct 11 · 06:09 UTC | +1,312.55 | -1,122.32 | -45.33 | -413.35 | +268.45 |
| Oct 11 · 06:09 UTC | +1,312.55 | -1,121.99 | -45.33 | -413.68 | +268.45 |
| Oct 11 · 06:09 UTC | +1,313.32 | -1,121.99 | -45.33 | -414.45 | +268.45 |
| Oct 11 · 06:09 UTC | +1,306.76 | -1,093.31 | -38.77 | -443.13 | +268.45 |
| Oct 11 · 06:09 UTC | +1,305.80 | -1,092.35 | -38.77 | -443.13 | +268.45 |
| Oct 11 · 06:09 UTC | +1,307.98 | -1,092.35 | -37.82 | -445.30 | +267.50 |
| Oct 11 · 06:09 UTC | +1,311.10 | -1,094.36 | -35.81 | -445.30 | +264.37 |
| Oct 11 · 06:09 UTC | +1,335.91 | -1,094.36 | -14.76 | -470.11 | +243.32 |
| Oct 11 · 06:09 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 06:09 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 06:09 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 07:12 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 07:12 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 07:12 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 07:12 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 07:12 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 07:12 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 08:15 UTC | +1,335.91 | -1,091.81 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 08:15 UTC | +1,342.40 | -1,098.29 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 08:15 UTC | +1,342.40 | -1,098.29 | -14.76 | -472.66 | +243.32 |
| Oct 11 · 08:15 UTC | +1,342.40 | -1,092.16 | -14.76 | -478.80 | +243.32 |
| Oct 11 · 08:15 UTC | +1,342.40 | -1,092.16 | -14.76 | -478.80 | +243.32 |
| Oct 11 · 08:15 UTC | +1,345.04 | -1,092.16 | -12.45 | -481.44 | +241.01 |
Accuracy & confidence
D1 / D3 / D5 are deadlines 24 / 72 / 120 hours before the observing day starts. Each score uses only finalized contests with all five eligible forecasts at that deadline, including byes and untraded entries. Initialization times may differ. No archive scores enter this table.
Confidence calibration · forecast versus observed frequency
Calibrated 80% forecasts should verify about 80% of the time. Small bins are noisy; the count stays visible.
| MODEL | PROBABILITY BIN | N | MEAN FORECAST | OBSERVED FREQUENCY |
|---|---|---|---|---|
| AIFS | 0–20% | 22 | 11.0% | 0.0% |
| AIFS | 20–40% | 28 | 29.5% | 3.6% |
| AIFS | 40–60% | 40 | 52.2% | 47.5% |
| AIFS | 60–80% | 40 | 72.2% | 80.0% |
| AIFS | 80–100% | 146 | 95.0% | 98.6% |
| EPS | 0–20% | 14 | 9.2% | 0.0% |
| EPS | 20–40% | 18 | 29.2% | 5.6% |
| EPS | 40–60% | 22 | 51.6% | 50.0% |
| EPS | 60–80% | 60 | 72.3% | 66.7% |
| EPS | 80–100% | 162 | 95.0% | 88.9% |
| WN2 | 0–20% | 12 | 10.1% | 16.7% |
| WN2 | 20–40% | 28 | 32.2% | 14.3% |
| WN2 | 40–60% | 37 | 50.2% | 35.1% |
| WN2 | 60–80% | 43 | 69.4% | 65.1% |
| WN2 | 80–100% | 156 | 95.2% | 95.5% |
| HRES | 0–20% | 34 | 7.9% | 0.0% |
| HRES | 20–40% | 33 | 28.8% | 33.3% |
| HRES | 40–60% | 42 | 51.1% | 71.4% |
| HRES | 60–80% | 36 | 68.8% | 80.6% |
| HRES | 80–100% | 131 | 94.0% | 96.2% |
| ICON | 0–20% | 64 | 6.0% | 14.1% |
| ICON | 20–40% | 30 | 30.4% | 63.3% |
| ICON | 40–60% | 31 | 49.1% | 74.2% |
| ICON | 60–80% | 34 | 68.4% | 82.4% |
| ICON | 80–100% | 117 | 94.8% | 100.0% |
The season slate
| CONTEST | TARGET DAY | LEAD | LOCK UTC | STATUS | LOCKED ENTRIES |
|---|---|---|---|---|---|
| KATL · High ≥ 82°F ↗ | 20260920 | D1 | Sep 19 · 05:00 UTC | Settled | 4 |
| KATL · High ≥ 82°F ↗ | 20260922 | D3 | Sep 19 · 05:00 UTC | Settled | 4 |
| KATL · High ≥ 81°F ↗ | 20260924 | D5 | Sep 19 · 05:00 UTC | Settled | 4 |
| KBOS · High ≥ 72°F ↗ | 20260920 | D1 | Sep 19 · 05:00 UTC | Settled | 4 |
| KBOS · High ≥ 71°F ↗ | 20260922 | D3 | Sep 19 · 05:00 UTC | Settled | 4 |
| KBOS · High ≥ 70°F ↗ | 20260924 | D5 | Sep 19 · 05:00 UTC | Settled | 4 |
| KLGA · High ≥ 76°F ↗ | 20260920 | D1 | Sep 19 · 05:00 UTC | Settled | 4 |
| KLGA · High ≥ 75°F ↗ | 20260922 | D3 | Sep 19 · 05:00 UTC | Settled | 4 |
| KLGA · High ≥ 74°F ↗ | 20260924 | D5 | Sep 19 · 05:00 UTC | Settled | 4 |
| KMIA · High ≥ 89°F ↗ | 20260920 | D1 | Sep 19 · 05:00 UTC | Settled | 4 |
| KMIA · High ≥ 89°F ↗ | 20260922 | D3 | Sep 19 · 05:00 UTC | Settled | 4 |
| KMIA · High ≥ 89°F ↗ | 20260924 | D5 | Sep 19 · 05:00 UTC | Settled | 4 |
| KDFW · High ≥ 88°F ↗ | 20260920 | D1 | Sep 19 · 06:00 UTC | Settled | 4 |
| KDFW · High ≥ 88°F ↗ | 20260922 | D3 | Sep 19 · 06:00 UTC | Settled | 4 |
| KDFW · High ≥ 87°F ↗ | 20260924 | D5 | Sep 19 · 06:00 UTC | Settled | 4 |
| KIAH · High ≥ 89°F ↗ | 20260920 | D1 | Sep 19 · 06:00 UTC | Settled | 4 |
| KIAH · High ≥ 89°F ↗ | 20260922 | D3 | Sep 19 · 06:00 UTC | Settled | 4 |
| KIAH · High ≥ 88°F ↗ | 20260924 | D5 | Sep 19 · 06:00 UTC | Settled | 4 |
| KMSP · High ≥ 71°F ↗ | 20260920 | D1 | Sep 19 · 06:00 UTC | Settled | 4 |
| KMSP · High ≥ 70°F ↗ | 20260922 | D3 | Sep 19 · 06:00 UTC | Settled | 4 |
| KMSP · High ≥ 69°F ↗ | 20260924 | D5 | Sep 19 · 06:00 UTC | Settled | 4 |
| KORD · High ≥ 74°F ↗ | 20260920 | D1 | Sep 19 · 06:00 UTC | Settled | 4 |
| KORD · High ≥ 73°F ↗ | 20260922 | D3 | Sep 19 · 06:00 UTC | Settled | 4 |
| KORD · High ≥ 72°F ↗ | 20260924 | D5 | Sep 19 · 06:00 UTC | Settled | 4 |
Page 16 of 39 · 936 contests. Filter by target day; all leads are shown.
How the models compete
A 28-day slate uses 12 airport stations at D1, D3, and D5. Each high-temperature line is the rounded 1991–2020 ACIS daily normal, frozen at publication. Opponents rotate; one of the five models has a bye in each contest.
The worker freezes the newest eligible initialization received and captured before the deadline, at most 36 hours old. It has 60 minutes to execute that cutoff; a missed window produces no retrospective trade. Missing data means no entry.
A probability gap of 4 percentage points qualifies. The higher forecast takes YES, the other NO, at their midpoint price. Gaps of 4 / 10 / 20 points request 25 / 50 / 100 contracts, reduced to fit both budgets. This measures forecast disagreement, not a separately estimated confidence score.
Each contract pays one credit, fully reserved by its two models. Each model can reserve at most 100 credits per station and target day across all leads, 500 per region and day, and 2,000 overall. Each lead gets at most one-third of the station-day limit. Headroom is allocated evenly across scheduled appearances at each cutoff before matching.
The observing day runs midnight to midnight in local standard time, including during daylight-saving time. Exact-station ACIS daily high readings must have a midnight observation hour and no flags. Results remain provisional for at least 72 hours after the day ends and 24 hours after the last changed reading; finalization requires a fresh reading. Corrections before finalization reset that clock. Unresolved contests void at seven days and refund both sides.
Contract finality is distinct from certified climate data. Later corrections are recorded for 30 days without changing finalized payouts or scores. No public accounts, deposits, or wagering controls are part of this league.
The publication behind the forecasts
858 verified publications · 539,375 forecast-to-publication links. Each link records when the evidence was observed; it never replaces the forecast’s original receipt time.
Oct 04 · 15:04 UTC · Verified Oct 04 · 15:12 UTC · 137 links
Publication and calibration evidence ↗ · Source menu · Frozen publication
Oct 04 · 14:52 UTC · Verified Oct 04 · 14:56 UTC · 126 links
Publication and calibration evidence ↗ · Source menu · Frozen publication
Oct 04 · 13:52 UTC · Verified Oct 04 · 14:08 UTC · 127 links
Publication and calibration evidence ↗ · Source menu · Frozen publication
The full menu hash, station forecast values, source cycles, publication time, and frozen projection must agree. Calibration artifact fingerprints are supporting evidence; model-release versions remain unknown and this season’s policy remains shadow-v1.
All publication evidence (JSON) ↗Every credit has a receipt
Season shadow-20260908-v1 · 1,989 verified journal records. Forecasts, opponent schedules, collateral, corrections, and transfers remain inspectable.
Download the journal and locked forecast receipts (JSON) ↗Latest SHA-256: 24f5e6442c9a9823160931c40991aa100ae9facd1d7f2807b5b0018b24b9024e