Man, this is right in my wheelhouse...one last thing I'll add for tonight is that I also feel like binning the respective subgroups la Nina's by intensity is too much finite detail that exacerbates the sample size issue. You can list omit the seasons that are awful intensity marches from your final forecast composite and/or take measures accentuate the better marches, like doubling them up in the final composite, etc. Binning them by intensity and structure singularly is enough...IOW, have weak, mod, strong, modoki, central-basin wide (hybrid) and east based composites...you don't need weak east based, mod east based, etc...you can just factor those details into your final forecast.
The cool thing about seasonal forecasts is that they are an art with an inherent level of subjectivity, which is a real paradox given that weather is a science. The data is the only element of the entire process that is objective because the interpretation, application format and presentation are not. The forecaster is the artist, and his/her brush is a methodology that is unique to him/her. It is a tool that has been honed, refined and individualized over time ro be applied to a hemispheric canvass of sorts. It's the synthesis of a plethora of variables that is applied and then expressed through your very own lens. And it is for that reason that it is the ultimate form of self expression and a real oasis of creativity within a see of science. This is why I have taken to it because I hate science, but I like to express myself through writing and this couples that with my winter obsession.