WxDuel Dashboard

WxDuel compares human weather forecasting accuracy against automated guidance and historical averages.

Every day, millions of people check the weather, but how accurate are those predictions? WxDuel was created to answer a simple question: Who really knows what the weather will do?

Alpha Test Mode: We are actively testing the automated forecasting players and the scoring engine. Please be advised that during this phase, statistical data and rankings may be volatile or inaccurate.

Current 7-Day Results

Last Updated: Jun 5, 12:30 PM EDT

# Model / User MAE
1 NB5 1.43°
2 NBM 1.51°
3 NWS 1.79°

Tip: Click on a name above to view their detailed performance audit.

View Full Analysis

Benchmark Players

WxDuel includes automated players. These "players" participate in every game to provide a consistent standard for comparison:

NBM

Powered by the National Blend of Models. This player uses the raw output from the 13Z NBM cycle to generate its prediction.

NWS

Powered by the National Weather Service digital forecast. The initial forecast is read at 13:30Z, with a final update retrieved at 18:30Z.

Normals

Powered by Historical Normals (1991-2020). This player uses historical averages, serving as the minimum score to beat.

GFS

Powered by the Global Forecast System. This player uses the raw output from the 00Z GFS model run (from the previous evening).

NAM

Powered by the North American Mesoscale (3km) model. A regional model using 12Z guidance for higher resolution over the US.

HRRR

Powered by the High-Resolution Rapid Refresh. Uses the 12Z run of this real-time model for short-term precision.

GEFS

Powered by the Global Ensemble Forecast System. Uses the 00Z ensemble mean to smooth out outliers and provide probabilistic guidance.

CMCE

Powered by the Canadian Meteorological Centre Ensemble. Uses the 00Z mean from the Canadian global ensemble system.

AIGFS

Powered by the AI Global Forecast System. The new deterministic AI model (00Z run).

AIGEFS

Powered by the AI Global Ensemble Forecast System. Uses the 00Z ensemble mean from the AI-driven ensemble system.

NB1 / NB9 (Extremes)

The 10th & 90th Percentiles of the 13Z NBM. These represent the available "cold" and "warm" outlier scenarios beyond the mean.

NB5 (Median)

The 50th Percentile of the 13Z NBM. This represents the statistical median, often serving as a highly accurate "middle ground" forecast.