On Dec 22, 10:44*am, "Norman" wrote:

Even if only model "performance" is considered you have to define what you mean
by that before you can make any proper comparison.

Here's one - 500mb anomaly correlation

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz5.html

Even then, it would be good to see how this statistics vary by weather
"type" - westerly, blocked etc.

Richard