|
| The resulting model features better anomaly and trend detection earlier in the season than the USDA’s official forecast. It achieves a high degree of accuracy even in very anomalous years. Running live through the 2016 growing season, the yield model predicted final yield for 2016 within 1.40% of the final yield by the end of August (final yield numbers are reported in January). In backtesting for the period 2001-2015, Gro’s model estimated national final yield with an average error of 2.69% by mid-August. When updated with 2016 data and backtested for the period 2001-2016, Gro’s model estimated national final yield with an average error of 2.44%. We were at least as good as, and in some cases better than, currently available private models available for purchase which provide no transparency into their methods to their users, let alone the broader public.
https://gro-intelligence.com/yield-model
Greetings All
The best access I have to corn yield model is the link above. Please take time to read through the narrative of the model, key take aways are how variable July USDA and Gro model estimate are and how much accurate the August WASDE and Gro's models are.... will be.
Once one reviews the accuracy of the August forcast compared to the USDA January Final Yield number, you understand the importance of that upcoming August WASDE number. Please note in years where we have drought conditions, WASDE August numbers errors on the high side...
Weather you agree with USDA number or not...the market will trade them...why mot accurate info the market can agree upon.
Best
| |
|
National average arguments
