I wish we had such a better database regarding winter storms. And by that I mean more organized. Not only that but obviously you have the discrepancies and lots of lost data from the mid-90's to the early 2000's.
Anyways if there was some sort of database where you can click on a winter season and acquire a list of every single storm/event (including totals) having such information I think could go a substantial way in long-range forecasting (pertaining to winter).
While there has been a ton of success with various methods out there relying heavily on analogs, or making a base comparison based on ENSO phase can only go far. I also wish there was a better breakdown of the many indices which exist instead of these "monthly averages". They're fine and good and all but they aren't going to highlight or address short-term phase changes and that is what is most important.
One example is the WPO which has a very strong correlation to the pattern configuration from the Pacific through North America and you can even argue it may even drive some of the northern hemisphere teleconnections. The WPO can also go through some wild swings on a seasonal basis.