He' laid it out the fundamentals of his new prediction idea about ten years ago in the Journal of Climate, although since then he's brought out a Snow Advance Index and with a fellow researcher at his firm has a Polar Vortex Index in the works. That index appears to be demonstrating predictive skill thirty days out.
He made the essential point that ENSO (El Nino Southern Oscillation) forecasts are skillful mostly for the West Coast in winter, and less so for the rest of the nation, whereas the nature of his polar vortex/snow advance projections speak for New England and Europe in winter. He also suggested that the global climate models have difficulty capturing heat exchanges at high altitudes at the poles.
The explanation for these "winter severity" forecasts is not simple, but here's a look at the warm Arctic/cold continents idea from NOAA, and here's an excellent story on Cohen from Wired, that teases out the various threads, and the connection to global warming.
Plus, an impressive graphic he gave to the magazine, showing how warming upsets the apple cart (the polar vortex) in the Arctic, leading to the famed Snowmaggedon of 2010.
The broad effects of global warming have been predicted for decades; now the shorter-term consequences are becoming easier to understand. Could this lead to the end of denial?