The high-bias in the forecast values reminds me of last winter. Could the bias be due to a poor coupling between the ocean model and the atmospheric model?
On the positive side, the high-bias may allow some fantasy storms to come up the coast. On the negative side, reality will be much more mundane.
Not that the AO forecast has much to say about it but I'm rooting for a warm November. My hunch is that the correlation between November temperatures and winter temperatures is negative - perhaps -0.02. I'll take any advantage I can get.
A depressing truth; the "smartest" forecast might be to use the ensemble with the most positive AO value.
Another depressing feature of the time series is that it goes negative about when the forecast skill goes to zero or at least to 0.16^2.