Are we allowed to further improve the models that we submitted in the forecast challenge for the overall prizes stage? Or do we have to use the same solutions that are now hopefully smoothly running in the forecasting stage?
I also made some model changes between hindcast and forecast challenges that went beyond retraining. Hope that was okay.
Hi @kurisu,
Thanks for your patience. The challenge organizers are still finalizing guidance to give regarding the upcoming final prize stage, and I will follow up when there is an update.
In the meantime, please be sure to track any changes you make to your model, and be sure to be able to produce the training code for your Forecast Stage model, which will be necessary regardless in the case that you’re a Forecast Stage finalist.