15 days out is way too long for any type of current-generation deterministic forecast. Just yesterday the GFS was predicting this for November 16.
Source:
https://www.tropicaltidbits.com/analysis/models/
The current, practical limit of predictability is about 9 or 10 days; beyond this time you need to use probabilistic forecasting techniques.
It may be possible to increase this limit a few days by improving accuracy of initial conditions, improving the model physics, and reducing its grid size. However, there is a theoretical limit of deterministic forecasting beyond which skillful forecasting is mathematically impossible. This theoretical limit is probably about 14 days: beyond that time even if you could measure the initial conditions with absolute precision and have no uncertainty in the model physics and have an arbitrarily small grid size it is still impossible to have a skillful forecast. This limit is established by the math; you can change it no more than you can make 2 plus 2 equal 5. Such is the magic of non-linear (chaotic) systems.
https://journals.ametsoc.org/view/jo...-18-0269.1.xml