Station-aligned Open-Meteo previous_day1

Open-Meteo Next-Day Forecast Brief

Priority cities only. Forecast update only; no odds/trading advice.

Priority cities only. Forecast update only; no odds/trading advice.

Summary: Current ready: 3 | Waiting for cutoff: 0 | Stale previous rows: 0 | Generated at: 7月1日 04:45

Core tier-1

City / stationTierForecast dateExpected dateForecast versionForecast maxRounded market bucketPolymarket outcomeUnitHistorical hit typeHistorical hit rateHistorical sample nDisplay statusFreshnessMelbourne triggerVisible byLast updated Melbourne
Paris / LFPBCore tier-12026-07-012026-07-01MAX_THROUGH_1726.2°C -> 26°C26°C26°CCsettlement-standard exact-number37.59%n=979AFTER_CUTOFFCURRENT_FORECAST_DATE7月1日 01:407月1日 01:457月1日 01:40
NYC / KLGACore tier-12026-06-302026-07-01MAX_THROUGH_1988.88°F -> 89°F89°F86°F or higherFstandardized range-bin27.07%n=979STALE_FORECAST_DATESTALE_FORECAST_DATE6月30日 09:406月30日 09:456月30日 09:40

New watchlist

City / stationTierForecast dateExpected dateForecast versionForecast maxRounded market bucketPolymarket outcomeUnitHistorical hit typeHistorical hit rateHistorical sample nDisplay statusFreshnessMelbourne triggerVisible byLast updated Melbourne
Lucknow / VILKNew watchlist2026-07-012026-07-02MAX_THROUGH_1731.9°C -> 32°C32°C32°CCsettlement-standard exact-number26.30%n=384STALE_FORECAST_DATESTALE_FORECAST_DATE6月30日 22:106月30日 22:156月30日 22:10
Denver / KBKFNew watchlist2026-06-302026-07-01MAX_THROUGH_1790.14°F -> 90°F90°F90°FFstandardized range-bin25.98%n=381STALE_FORECAST_DATESTALE_FORECAST_DATE6月30日 09:406月30日 09:456月30日 09:40
Mexico City / MMMXNew watchlist2026-06-302026-07-01MAX_THROUGH_1723.0°C -> 23°C23°C23°CCsettlement-standard exact-number33.07%n=381STALE_FORECAST_DATESTALE_FORECAST_DATE6月30日 09:406月30日 09:456月30日 09:40

Current ready rows

Ready rows are current forecast_date rows whose cutoff has passed.

City / stationDateVersionForecastRoundedOutcomeUpdatedStatus
London / EGLC2026-07-01MAX_THROUGH_1924.5°C25°C25°C7月1日 04:40AFTER_CUTOFF
Madrid / LEMD2026-07-01MAX_THROUGH_1835.9°C36°C36°C7月1日 02:40AFTER_CUTOFF
Paris / LFPB2026-07-01MAX_THROUGH_1726.2°C26°C26°C7月1日 01:40AFTER_CUTOFF

Waiting for cutoff / upcoming rows

Waiting rows are reserved for future display states not covered by current ready or stale/reference rows.

Stale previous rows

Stale/reference rows are current latest forecast rows shown before their cutoff. Forecast date means the target date whose maximum temperature this row forecasts. They are not treated as current ready rows.

Metadata and caveats
  • generated_at_melbourne: 2026-07-01T04:45:30+10:00
  • forecast_date range: 2026-06-21 to 2026-07-01
  • source: station-aligned Open-Meteo previous_day1
  • caveat: forecast update only, no odds/trading advice
  • caveat: cutoff_status is based on configured Melbourne trigger time
  • caveat: Rows marked STALE_FORECAST_DATE are for an older forecast_date than the current expected station-local target date. They are shown for audit visibility and should not be treated as current forecasts.
  • caveat: Rounded market bucket is the model forecast aligned to the market's whole-degree precision. Polymarket outcome is the actual displayed market bin/range containing that whole-degree value.
  • caveat: Historical hit rate is research-only: exact-number matching for London/Madrid/Paris and standardized 2°F bucket simulation for Austin/NYC. Not a probability, odds, or trading signal.
  • caveat: Historical hit rate is sourced from the displayed-city settlement-standard research summary when available and shown as one long-sample backtest reference per city.
  • caveat: NYC / KLGA uses standardized settlement-standard even-lower/odd-upper 2°F range-bin research for the displayed Historical hit rate; open-ended market tails are ignored.
  • caveat: The prior NYC 57.53% bucket metric was tail-only and is no longer used for current range-bin display.
  • caveat: London / EGLC displays settlement-standard exact-number research and remains the strongest clean/near-complete exact-number city in the current long-sample summary.
  • caveat: London MAX_THROUGH_19 range-bin recompute files are retained only as historical diagnostic artifacts and are not used for the active London frontend metrics.
  • caveat: Austin / KAUS uses standardized settlement-standard even-lower/odd-upper 2°F range-bin research for the displayed Historical hit rate; open-ended market tails are ignored.
  • caveat: Denver / KBKF uses standardized settlement-standard even-lower/odd-upper 2°F range-bin research for the displayed Historical hit rate; missing observed dates remain 2025-08-30, 2025-08-31, and 2025-09-01.
  • caveat: Paris / LFPB frontend display uses Paris-Le Bourget / LFPB as the current-market settlement standard for the Historical hit rate. Older LFPG evidence is retained only as a historical/source-drift diagnostic; no station mapping or forecast definition changes are made.
  • caveat: Mexico City / MMMX is mixed-provenance near-complete settlement-standard exact-number research; the unresolved 2026-06-19 Open-Meteo source mismatch was preserved and not overwritten.
  • caveat: Madrid / LEMD uses an early forecast cutoff (MAX_THROUGH_18): the displayed Historical hit rate is computed from the station-local 00–18 forecast window. The settlement target remains the full-day observed maximum; this early cutoff was selected because the cutoff feasibility audit found minimal historical accuracy loss versus FULL_DAY_MAX. The Polymarket market does not settle at 18:00.
  • note: This brief displays core tier-1 cities and new watchlist cities. Backend CSV/JSON outputs continue to refresh all configured live cities.
  • note: Core tier-1 cities are established live forecast monitors.
  • note: New watchlist cities are promoted from the new-city discovery audit and should be monitored separately before being treated as core.
  • note: Backend CSV/JSON outputs continue to refresh all configured live cities.
Historical research only · display-only

Latest audit-closed research status

Latest local closeouts are consolidated here as compact display text. These are historical descriptive sample results only: not forward-looking predictions, not probabilities, not active, and not frontend/live approved.

Austin — corrected forecast=Top1

n=41 · official-fee historical net +3.81529

  • Rule: corrected forecast = market Top1.
  • Average historical net per entry: +0.093056.
  • Status: audit-closed positive but sample-limited.
  • Interpretation: priority forward-only strategy tracking candidate; sample-limited and not frontend/live active.

Paris — MARKET_TOP1_UNCONDITIONAL

n=118 · official-fee historical net +4.55660

  • Average historical net per entry: +0.038615.
  • Status: market-calibration forward-track candidate, not forecast signal.
  • Interpretation: historical market-baseline anomaly only; not forecast resonance.

Milan — MARKET_RANK2_UNCONDITIONAL

n=89 · official-fee historical net +2.65565

  • Status: offline watch-only / exploratory overfit risk.
  • Caveats: train 70% negative -0.79177; June/latest month drives result; removing June flips negative -0.20507.
  • Top 5 winner removal flips negative -1.07823; bootstrap interval includes zero.
  • Interpretation: watch-only, not robust enough for active display/live use.

Helsinki — watch-only

Quote/rank backfill completed

  • Status: watch-only.
  • No conservative strategy shortlist.
  • No strategy forward-track candidate.

Warsaw — NEEDS_DATA_FIRST

CHECKPOINT_REPAIR_FAILED_SOURCE_LIMITED

  • Token/outcome mappings: 94 market dates / 1045 rows.
  • checkpoint_utc missing for all mapped dates: 0/94 usable.
  • Do not run Telonex probe/backfill/strategy until a defensible checkpoint source is found.

London — odds strategy excluded

n=244 · official-fee historical net -11.73382

  • Average historical net per entry: -0.048089.
  • Useful for forecast-accuracy monitoring only.
  • Not odds strategy tracking.

NYC / Madrid — not promoted

Weak candidates remain below promotion threshold

  • NYC: weak/source-limited candidate only.
  • Madrid: weak/outlier-driven candidate only.
  • Neither is promoted to frontend/live use.

Top1 / Top2 labels in this section refer to highest-ask outcome ranking in historical samples.

Retired / superseded: Austin tail-only, Austin KILL, Paris forecast resonance, NYC n=4-only, the old handoff net-value table, and 0.04/share as the primary fee model.

This section is historical research only, display-only, not active, not frontend/live approved, and not trading advice.

Archived historical appendix — Telonex Ask-only Top2 Resonance (superseded)

This appendix is retained for audit history only. It is superseded by the audit-closed research status above and should not be treated as the current strategy reference.

Archived research · display only

Offline Research Appendix: Telonex Ask-only Top2 Resonance

Archived offline research closeout. This superseded appendix does not update probabilities or forecasts.

Primary pooled result (non-Madrid)

  • Included cities: Austin, London, NYC, Paris
  • Excluded from primary pooled result: Madrid
  • Forecast-only exact: 123 / 269 = 45.72%
  • Resonance exact: 111 / 210 = 52.86%
  • Historical lift: +7.14 percentage points
  • Settlement in ask-only Top2: 206 / 269 = 76.58%
  • Resonance average forecast ask: 0.3841
  • Average decimal quote from ask: 2.7170

City-level summary

CityComplete datesAsk-rankableForecast exactResonance exactNotes
Austin847241 / 70 = 58.57%39 / 55 = 70.91%Available 84d, not full 100d
London1006928 / 67 = 41.79%25 / 55 = 45.45%—
NYC1006824 / 66 = 36.36%20 / 44 = 45.45%—
Paris1006630 / 66 = 45.45%27 / 56 = 48.21%—

Madrid appendix

  • Madrid is not part of the primary pooled result.
  • Standalone rescued sample, not 100d.
  • Complete dates: 89
  • Ask-only rankable: 86
  • Forecast exact: 31 / 86 = 36.05%
  • Resonance exact: 31 / 77 = 40.26%
  • Settlement in ask-only Top2: 69 / 86 = 80.23%

Definitions and caveats

  • Resonance means the forecast outcome is inside the ask-only full-market Top2 before checkpoint.
  • Ask-only Top2 is a research proxy.
  • Midpoint is diagnostic only because all-outcome midpoint coverage is sparse.
  • This is archived offline research only.
  • This does not update probabilities or forecasts.
  • This is superseded by the audit-closed research status.