What the market asked
The Polymarket event was straightforward in framing and strict in resolution: what would be the highest temperature in London on 25 August 2026? Settlement depended on a single official source and rule set—the highest whole-degree Celsius value in the NOAA/NWS WRH Time Series Viewer Temp column for station EGLC (London City Airport), metric units, with the first observation of the 26th locking the calendar day. The market was structured around discrete temperature outcomes rather than a simple binary, so the useful question for a council was which bucket was most probable and at what strength.
What the council concluded
Report date was 24 August 2026. Participating models were Grok, Claude, and GPT, with Grok as chairman. The council’s graded pick—the most-likely outcome—was: “Will the highest temperature in London be 24°C on August 25?” at 26% probability. Overall council action was logged as BUY_YES with a 20% probability and 58% confidence. The quoted reasoning stated:
“Council recommends BUY_YES (YES). Probability: 20%, Confidence: 58%. Resolution requires the single highest whole-degree Celsius value reported in the NOAA/NWS WRH Time Series Viewer Temp column for station EGLC (London City Airport) on 25 August 2026 (metric units); the first observation of the 26th locks the day.”
In plain terms, the panel treated 24°C as the mode of the distribution while still expressing limited conviction: a one-in-four-ish chance on the top bin, and a lower headline probability on the YES framing they attached to the trade recommendation. Weather markets one day out are often wide; the council did not pretend otherwise.
How the market was priced relative to the pick
The council’s own numbers are the only pricing signal in the record used here. It put 26% on 24°C as most likely and 20% on the BUY_YES action, with confidence at 58%. That is not a high-edge, high-certainty stance. It is a modal forecast inside a multi-outcome temperature ladder—consistent with late-summer London variability and with the hard EGLC/NOAA definition that leaves little room for narrative after the observations post.
What actually happened
The winning outcome was exactly the graded pick: the highest temperature in London on 25 August 2026 resolved to 24°C under the stated EGLC rule. council_was_correct is true on the most_likely grade basis. The Brier score for the pick was 0.5476, which matches a 26% forecast on an event that occurred: (0.26 − 1)² = 0.5476. Correct direction, modest probability, therefore a middling probabilistic score rather than a sharp win.
Takeaway
This was a clean resolution path and a correct mode call. The council named 24°C as most likely, the station data agreed, and the grade followed. Honesty still requires noting the calibration: 26% on the winner is better than missing the bin entirely, but it is not strong foresight, and the accompanying BUY_YES probability (20%) and confidence (58%) already signaled restraint. For weather markets tied to a single METAR-style series, process discipline—cite the station, the column, and the lock time—matters as much as the percentage. Here that discipline lined up with the outcome; the Brier number simply records that the council did not overclaim certainty the day before.
AI-generated analysis for informational purposes only. Not financial advice.