As far as predicting what will happen in 2050, more skill needs to be shown on shorter time frames first.
I believe you have that backwards. If you could show any degree of skill at all on a one-month time frame you could base a profitable company on that ability.
It’s the long-term predictions that we have a better chance with. I can predict July will be warmer than February for Northern California with my eyes closed, but I can’t predict whether next week will be warmer than this week without at least a little extispicy.