i find the anomaly charts are pretty stable, for a period of 'x' days then they have a re-alignment .... abit like correcting yourself whilst driving through fog at night. the charts you posted shows just that. what i dont think ive ever seen with them, is a total switch - replacing high with low/positive with negative.
ive found that when the anomalies stand alone, they are still proven accurate more times then not. that has happened several times over the winter period - remember those stunning easterlies? noaa's said no... they were correct and just yesterday the gfs and ecm op runs backed off from their over progressive eastward track . the noaa's said 'no' and they again were proven accurate.
as i see it, its about % 's... which suite is most accurate for the timespan in question, NO one model is infalible, the noaa's are not always right. but they do appear to be more accurate , no, they ARE more accurate then the ops for the timeframe they pertain to.