Warning: count(): Parameter must be an array or an object that implements Countable in /home/tglsjadd/public_html/pghsnowday/wp/wp-includes/post-template.php on line 284

Warning: count(): Parameter must be an array or an object that implements Countable in /home/tglsjadd/public_html/pghsnowday/wp/wp-includes/post-template.php on line 284

Warning: count(): Parameter must be an array or an object that implements Countable in /home/tglsjadd/public_html/pghsnowday/wp/wp-includes/post-template.php on line 284

Warning: count(): Parameter must be an array or an object that implements Countable in /home/tglsjadd/public_html/pghsnowday/wp/wp-includes/post-template.php on line 284

Forecaster Ratings

At PGH Snow Day, we have been collecting and analyzing forecasts from KDKA, WTAE, WPXI, AccuWeather, Intellicast, NWS Pittsburgh, the NWS Digital Forecast Model, and The Weather Channel since the 2015-2016 winter. This data is housed in our Winter Archive.

These forecasts are stored in our roundups, then analyzed and assigned accuracy scores after the final storm total snowfall is reported by the National Weather Service. At the conclusion of the winter, we average all individual scores into a final season accuracy score. This season score is then averaged with all previous season scores, and this overall accuracy score is used to calculate the ratings shown on this page.

The method by which individual accuracy scores are calculated is as follows:

  • An average forecast amount is calculated by averaging the forecast range. For example, a forecast of 1-3″ has an average forecast amount of 2″.
  • The actual snowfall amount is subtracted from the forecast amount to determine forecast error.
  • The forecast error is subtracted from 10 to determine the accuracy score.
    • Accuracy score cannot be lower than 9.0 if the actual snowfall was within the forecast range.
    • An outlet forecasting 0″ is automatically assigned a maximum score of 5 when actual snowfall is 1″ or more.

The following are our Forecaster Ratings as of November 4th, 2019. Be sure to hover over individual points on the graphs, they are interactive! (Note: these graphs are continuing to be updated to include the 2018-19 forecasts. All accuracy scores are up to date.)


KDKA (CBS Pittsburgh) B
CBS Pittsburgh has an overall accuracy score of 83.31%.

We have collected 27 forecasts from KDKA. Their most accurate winter was the 2018-19 season, where they recorded an averaged accuracy score of 90.50%

 


WTAE (ABC Pittsburgh) B
WTAE has an overall accuracy score of 88.23%.

We have collected 29 forecasts from WTAE. Their most accurate winter was the 2015-16 season, where they recorded an averaged accuracy score of 92.28%

 


WPXI (NBC Pittsburgh) B
WPXI has an overall accuracy score of 85.61%.

We have collected 28 forecasts from WPXI. Their most accurate winter was the 2015-16 season, where they recorded an averaged accuracy score of 89%

 


National Weather Service (Pittsburgh) B
The National Weather Service Pittsburgh office has an accuracy score of 86.44%.

We have collected 28 forecasts from the Pittsburgh Office of the NWS. Their most accurate winter was the 2018-19 season, where they recorded an averaged accuracy score of 92%.

 


AccuWeather B

AccuWeather has an accuracy grade of 87.04%.

We have collected 29 forecasts from AccuWeather. Their most accurate winter was the 2018-19 season, where they recorded an averaged accuracy score of 89.83%.

 


Intellicast B
Intellicast’s Snowfall Model has an accuracy grade of 85.19%.

While Intellicast was still publishing, we collected 25 forecasts from Intellicast. In the 2015-16 season, they averaged an accuracy score of 9.28/10, or 92.78%. In the 2016-17 season, they averaged an accuracy score of 89.50%. In the 2017-18 season, they averaged an accuracy score of 82.75%. In the 2018-19 season, they averaged an accuracy score of 75.75%. Over the four season we tracked, they recorded an average accuracy score of 85.19%.

 


National Weather Service (National Model) C
The National Weather Service’s National Snowfall Model had an accuracy grade of 79.24%.

We collected 21 forecasts from the NWS Digital Model from 2015 until 2018. In the 2015-16 season, they averaged an accuracy score of 7.8/10, or 78.00%. In the 2016-17 season, they averaged an accuracy score of 90.38%. In the 2017-18 season, they averaged an accuracy score of 69.33%. Overall, they had an accuracy score of 79.24%.

 


The Weather Channel B
The Weather Channel has an accuracy grade of 88.47%.

We have collected 18 forecasts from The Weather Channel. Their most accurate winter was the 2016-17 season, where they recorded an averaged accuracy score of 90.25%

 


 

Is there a forecast not listed above that you think we should be tracking? Let us know: pghsnowday@gmail.com

Note: We understand that predicting the weather, especially in the winter, is extremely difficult. Our ratings are meant as a way to analyze the success of individual forecasts and models, and are not meant to discourage trust in forecasts. Please heed all weather watches and warnings and listen to the forecasts of these professional meteorologists.