I think you're looking at them backwards. The 0Z was the top chart, the 6Z was the bottom one. The means are as follows, from Jan 22 to Jan 27, midnight each day (to nearest whole degree).
0Z ... 6Z
-3 .... -4
-3 .... -3
-2 .... -3
-2 .... -2
-1 .... -2
-1 .... -4
Nope. The entire post of yours that I quoted didn't mention 0z
"GEFS now out for the 06Z and have trended a fair bit colder in the mid term for London, Jan 22-27, with the colder cluster predominating now. Control run is fairly frigid."
Yes they are better than the 0z. I thought you meant the 06z was trending lower for the period quoted not compared to the 0z.
-3
-5
-4
-2
-3
-4
Trending milder then a colder one at the end I would say, but in the scope of variability I would say remaining between -1 and -5C