It appears that the extended forecast is still struggling considerably with the blocking pattern in NW Europe. Shown below are the last 4 runs. In each one of them, I have circled the week commencing 04.10.2021. In the run of 23.09. an almost equal chance of blocking and +NAO was suggested. In th two subsequent runs, 27.09 and 30.09, the probability tilted strongly and clearly towards the +NAO scenario. And then, coming on the run from yday (04.10), all the +NAO clusters disappeared and blocking is now the most likely scenario. 
The question of course is how could a forecaster rely on a sub-seasonal forecast (actually the issue is here already in the D+7 to D+15) with such big volatility? But the most useful would be a brainstorming discussion on what could be the likely reasons apart from the obvious chaotic nature of the atmosphere.