Highest temperature in Austin on Oct 21, 2026: PBE moves “89° to 90°” up to 32% from 5%
New forecast guidance reached the model between 14:02 UTC and 17:00 UTC on Oct 20. Same model version, later data cutoff: here is exactly which inputs changed.
Photo: Quintin Soloviev · CC BY 4.0 · Wikimedia Commons
89° to 90° · pbe-weather-maxtemp@1.1.0 · 14:02 UTC → 17:00 UTC
Highest temperature in Austin on Oct 21, 2026?
89° to 90° · PRE-WINDOW
Quick read
- pbe-weather-maxtemp@1.1.0 moved “89° to 90°” from 5% to 32% (+27 pts).
- Later data cutoff: Oct 20 11:00Z → Oct 20 17:00Z; same model version, so the move reflects inputs, not a model change.
- Primary inputs changed: GFS MOS forecast high, National Blend of Models forecast high.
- The Kalshi mid captured with each snapshot went from 8% to 3%.
What changed in the inputs
Between the snapshot captured at Oct 20 14:02Z (data cutoff Oct 20 11:00Z) and the one at Oct 20 17:00Z (cutoff Oct 20 17:00Z), the same model version — pbe-weather-maxtemp@1.1.0 — moved “89° to 90°” from 5% to 32%. The model inputs that changed: GFS MOS forecast high from 82°F to 85°F; National Blend of Models forecast high from 85°F to 88°F.
| Input | Before | After | Primary |
|---|---|---|---|
| Guidance lead time | 24h | 18h | — |
| GFS MOS forecast high | 82°F | 85°F | yes |
| National Blend of Models forecast high | 85°F | 88°F | yes |
This page reports which stored inputs differ between two immutable forecasts. It does not claim why the underlying weather guidance changed.
The forecast path
Data table
| Time (UTC) | Source | Value % |
|---|---|---|
| Oct 20 14:02Z | PBE pbe-weather-maxtemp@1.1.0 | 5 |
| Oct 20 14:02Z | Kalshi mid | 8 |
| Oct 20 14:04Z | Kalshi mid | 13 |
| Oct 20 14:12Z | Kalshi mid | 9 |
| Oct 20 14:14Z | Kalshi mid | 6 |
| Oct 20 14:36Z | Kalshi mid | 4 |
| Oct 20 15:00Z | Kalshi mid | 4 |
| Oct 20 15:06Z | Kalshi mid | 3 |
| Oct 20 16:00Z | Kalshi mid | 3 |
| Oct 20 17:00Z | PBE pbe-weather-maxtemp@1.1.0 | 32 |
| Oct 20 17:00Z | Kalshi mid | 3 |
The Kalshi mid captured with each snapshot went from 8% to 3%.
The two snapshots
| Earlier | Later | |
|---|---|---|
| Captured | Oct 20 14:02Z | Oct 20 17:00Z |
| Data cutoff | Oct 20 11:00Z | Oct 20 17:00Z |
| Guidance runs | NWS GFS MOS 2026-10-20T06Z, NWS National Blend of Models 2026-10-20T06Z | NWS GFS MOS 2026-10-20T12Z, NWS National Blend of Models 2026-10-20T12Z |
| PBE | 5% | 32% |
| Market at capture | 8% | 3% |
Evidence & method
What this story is based on. Every value above was read from these immutable records.
- Model
- pbe-weather-maxtemp@1.1.0 · RESEARCH
- Snapshot
- 9a26fc10-03c9-46ae-aba1-499b9c0f9906
- Compared with
- 6f2b8dd4-cf09-4a22-b9ff-0478f63acff8
- Captured
- Oct 20 17:00Z
- Data cutoff
- Oct 20 17:00Z
- Market observation
- 3% · captured with the snapshot
- Source families
- NOAA/NWS via Iowa Environmental Mesonet archive · NOAA Regional Climate Centers ACIS
- Resolves on
- The Weather Company (weather.com/kalshi) — The Weather Company daily maximum temperature for CLIAUS (weather.com/kalshi)
- Independent check
- NWS Daily Climate Report CLIAUS (WFO EWX); NOAA RCC-ACIS USW00013904
- Measurement
- Maximum temperature for the climate day 2026-10-21, 00:00-24:00 local standard time (UTC-6)
- Rounding
- Whole degrees Fahrenheit as reported; preliminary values may differ by rounding/conversion
- Exceptions
- Preliminary data may be revised; the exchange may hold expiration until a non-erroneous revision, else last fair price
- Pre-window only; no current-conditions input
- Guidance can run cold/hot in unusual regimes (see LA diagnostic, 2026-10-04)
Weather · daily high — inputs: National Blend of Models day max (GFS MOS fallback) + station empirical guidance-error tables. This model on the research board →
Market prices are a benchmark only and never enter a PropBetEdge model. Research-stage probabilities; not advice.