I think surprisingly cold would require significantly negative anomalies.
Anomalies are relative. There can be a positive anomaly for an early reference period, AND a negative anomaly for a recent reference period.
Also one could make the case that it's "surprising" that there is still (multi-year) ice in the arctic, long after the 'predicted' date of its end in 2013. Surprise is relative to expectation, and while "insignificant anomalies" is a perfectly reasonable expectation to have, it's still a subjective one.
Personally, my subjective expectation was for milder temperatures, because I thought the low September ice concentration would enhance heat flux. That effect slowly died off from mid October, so I'm surprised.
Perhaps your expectations were wiser than mine :-)