I understand your thinking, Gandalf. But these are numerical weather predictions. Any difference in the numbers on the small scale (the short waves as they appear on the charts we see) is going to be multiplied cubically in a 3-dimensional atmosphere. Thus the continent-wide background pattern is also likely to differ.

Mind you, I admit that does sound like a pretty compelling argument for ignoring these charts any more than 5 days ahead, never mind 5 months. 

How the algorithm works is not really the issue. Gandalf is right in that a monthly anomoly chart shouldn't swing as wildly as the daily prediction so far ahead because over a month the variation should iron out so long as the broad scale pattern is modelled fairly accurately.
In that respect the huge swings we are seeing on CFS renders it useless in setting even trends for four or five months ahead.
I for one won't be fooled.

Apologies if this is blunt, but they haven't.

If by 'showing high latitude blocking' you mean on selected days for selected months then you're probably right, but there is as far as I can see no trend and no pattern from one run to the next. At this range it is generally accepted that the CFS is as useful as a fart in a spacesuit.
Happy to see evidence that says otherwise though.

The thing is though Lanky to look at the overall trend and there are for more runs showing things similar to bottom chart for

January than there are showing something similar to the top.

For once I would welcome a 'mild' winter, the thought of 3 months of freezing in my cashmere wool coat which is meant to be warming is not inviting, as has been the form horse post-2010.

I am quite a fan (perhaps absurd in many eyes) of the Atlantic all-day rain event/storm... I think it's the watching it from the inside.  

I sincerely hope Gav's video doesn't come off.

The 'point' surely is to work at the problem of long range forecasting until such time as they *are* of value. OK, we may never get there but we certainly won't if we don't even try.

60 years ago 5 day forecasts were pretty useless, now they are reasonably accurate.

And that was the point Gav was making as well his video that started this thread. He also warned not to take them too seriously !

My point is that there is no conceivable way using recognisable inputs such as SST's, Arctic Ice, historical matches ete etc that the same month's average pressure deviation could ever change like that over 2 consecutive day's runs.

Without knowing their  list of inputs I can only guess, but it looks to me as though they have taken one of the numerical models and instead of running it forward for 0-120 hours, they have gone into FI territory and out the other side by nearly 6 months where the output is 99.99% noise generated by the laws of chaos theory.

Absolutely and there is nothing wrong with long-term experimentation with the aim of refinement and that is good and proper science, but in the same vein it is an undeniable fact that forecast models attempting to predict trends and global anomolies beyond say a month are not proven at this time, and knowing what little I know about the science of atmospheric physics it is a small miracle that we can get anywhere over 80% verification at 5 days, and maybe 60% at best at 10 days. I haven't seen any verification stats for the fabled EC32 but I predict it'll be 50% at very best.
What confidence therefore should we have at 120 days? <5% would be generous.

 Me too! Here`s hoping for lots and lots of rain,we now need it!!

Plenty of snow please .......................let it melt

You make a valid point Lanky, The CFS never the less did a pretty good job of picking up on the trend  of our winter getting colder from mid Jan onwards last year. and held true to a very cold march from this time last year onwards. I know this because I viewed every 9 month run  every day from late June til the start of Winter.

The CFS V2 has consistently been gunning for a HLB dominated January since May. Trust me on this because I've been tracking 80-90% of the updates since May time. I have a myriad of charts saved to my PC as proof. I think people are failing to understand how the CFS works. You don't take each update in isolation, you have to average the runs in order to see the bigger picture. You have to hold a week or two of runs and then assess the general outcome as portrayed by these runs. About 70% of runs for January (since May) have plumped for a very consistent scenario of higher than average heights around Iceland and Greenland. This isn't pie in the sky- it's fact. Each time this blocking scenario is shown, the synoptic set up is virtually identical to the last time it was shown.

Obviously it's only September and the CFS may be barking up the wrong tree. However, there is absolutely NO doubt what the model has been insisting on from May until now.

Quite so Crewe Cold. The way I look at it is that the cfs four times dAily nine month runs are essentially just a great big ensemble. Just as in an ensemble forecast  the starting data are tweeked so it is with the four daily runs out to nine months, just on a longer time scale. The fact that at the moment some 70 % or more of runs are going for high latitude blocking I think is of some statitical significance. Especially bearing in mind the consistent and similar percentage of runs from July onwards last year that went for a mid January start to cold and a very wintery March for the UK

I'm open to the arguments, but to be statistically significant as you put it I wondered where the 70% figure is derived from. Was it 70% of all runs in a specified period, and if so what period? If it is, as you say, '70% of runs at the moment' is that 70% of the four daily runs on a specific day (difficult as 3 of the 4 would be 75%), or over a few days or even the last week or so? I see 'end of May' has been mentioned.
This information would be useful as I am trying to seperate sound analysis from subjective interpretation.

70% is a rough figure which covers the period from mid May to present day. Obviously it's not surprising that the CFS throws up HP to south scenarios occasionally (like it has this morning) as this lies within the 30% updates which favour a non blocked scenario. Although I don't have the exact % of runs going for a blocked winter within the test period, I have no reason to exaggerate or make it up.