It will probably be a cold winter. Be prepared. The temperature trend is reliably predicted to get colder on average each year through at least 2020.
Professor Valentina Zharkova uses a very predictable, natural, observable occilator which has been observed for millennia (the sun and its sunspots) to create an algorithm (she calls “dynamo”) which models that highly reproducible solar cycle. Then she compares dynamo to the principal components of the various cycles in other variables (e.g. cosmic radiation) and events (e.g. Maunder temperature minimum) and determines statistical correlations between dynamo and each variable using standard statistical techniques (PCA.) When a strong correlation with the highly reproducible solar sunspot cycle is discovered, she then compares her dynamo algorithm to deep historical data for each variable to further confirm whether or not the correlation fits. Finally, if there is a strong correlation fit with the long term historical data, she uses the same dynamo algorithm to forecast future temperature cycles and solar events.
Fully understanding the cause of the sunspots, solar radiance and cosmic rays is not necessary because the solar occilator is highly reproducible and predictable.
By analogy, no one knows all the factors and cause/effect variables involved in amplification of radio waves, but many billions of radio amplifiers are used every day (e.g. your wifi, TV, radio and cell phone) and they are extremely reliable.
Video of her presentation followed by Q&A: https://www.youtube.com/watch?time_continue=2&v=M_yqIj38UmY
Bio of Professor Zharkova: https://www.northumbria.ac.uk/about-us/our-staff/z/professor-valentina-zharkova/