I think in general people over-estimate the predictive power of an H5 pattern on surface outcomes. Especially for a smoothed out mean. We can only say, x features are correlated with Y outcome, but exact placement, strength, interaction between them is an inexact science at best, even if the output is largely correct.
Snow maps help to show where cold and moisture is linking up broadly, but are prone to random variation. The addition of the 100+ members of the new EPS will be a big help in this respect, assuming they are sufficiently perturbed. Temp and moisture anomaly maps together can get you a little more detail on how it's all coming together, but don't show how individual waves may potentially play out. All this together is what makes d5+ forecasts a complete crapshoot.