Based on GFS in the last few days you would have probably concluded:

A mixed outlook with colder conditions at times

Based on ECM deterministic:

A cold outlook with a moderate to high chance of it turning very cold with disruptive snow becoming an increasing risk

You've only got to look at the Met Office long ranges (first section) in recent days to see that. The reason Mete Group are streets ahead this winter is because they aren't stake holders in a particular model and so haven't over weighted it in their analysis. Going off at a tangent, but it provides a good reason for separating data from forecasting as they do in the US.

The biggest change this morning is the failure of the increasing heights over Scandinavia connecting with the ringing Auntie Azores high, being cut off by the trailing trough between Greenland and Iceland, potentially leading to short wave formation and cutting through the forming block, so neatly described by last night's ECMF. 240 coming out - so we'll see....

Indeed, I’m just listening to the forecast 

As opposed to the Griceland heights of the 12z ECM compared to the PV engulfed 0z for the same area, you mean?

PS: Sorry I take it all back. ECM 00z 240 is completely consistent with ECM 12z 240. Look at all those oranges again on the upper half of the plot. 

It runs 4 times a day CC, though I think (could be wrong) the 06z and 18z runs only go out as far as 144hrs. Handy for fine tuning shorter term patterns. 

Some people won't listen though Brian.

People see what they want to see. 

ECMF 240 forms a banana shaped high from the Azores to Scandinavia, right on top of us, cold in the south. Lovely dry frosty weather - hopefully sunny too. Could be glorious if no fog. Could be hoar frost if freezing fog. Looks increasingly dry. 

I've never seen a 6z or 18z ECM 

Come on, you should know by now that a single poor set of model runs negates any number of decent runs and means winter is over! 😉

Me neither but apparently they exist and run out to 90 hours. Official documentation is here:

https://www.ecmwf.int/en/forecasts/datasets/set-i

Even regards this upcoming 'slider low', I think the GFS picked up on its general track quicker than the ECMWF if my memory serves correct. 

No I have said clearly what I mean. As a rule the ECM (with only a few examples) transitions from one scenario to another, often the GFS goes into lucky dip mode with erratic jumps. I don’t mind whether you believe me or not and I think the GFS has some strong points in picking up patterns. It noted today’s incoming wintry weather a couple of weeks ago! It’s in the middle bit when the time period gets closer (days 6 to ten) that the erratic nature becomes evident. As I say I don’t really mind if you or indeed Brian believes me.

Edit: Sorry Rob, I’m being silly.  (Although I would like a direct reply from Moomin answering my question as his logical seems ridiculously flawed “

Yes, I agree. It is the one I referred to in a previous post earlier this morning. 

Wow the ECM has had 7 shades of the proverbial kicked out of it this morning was not expecting that. Kudos to the GFS if proved right. ECM does have big issues in overdoing HLB in our part of the world and its screwed up again it would seem. Things can change but if today's output is accurate ECMs 10,15 and 46 day products will have all proved pretty useless for our part of the NH.

It'd be funny if it wasn't usually the case. 

You can get short term local data from the 06z & 18z runs on the 'weather.us' site. I have it set to my location but you can put in yours:

https://meteologix.com/ie/forecast/2964180-galway/essentials/rapid-euro

This particular meteogram (Rapid ECMWF/Global EURO HD) updates 4 times a day. Currently shows the 00z forecast but will update to the 06z run during the afternoon and 18z later tonight. 

Well yes that's also a very good point. The 46 dayer is obviously ensemble based but you have to suspect it has the same shortcomings. 

I’m not sure that’s true Brian - it’s clear the Met Office use all the data at their disposal and will bin their own model and ensembles if they think they’re barking up the wrong tree. 

MeteoGroup tend to use ECM as their primary model I think, but the same applies. 

Forecasters are savvy enough to never solely put weight on a single model, especially the pros. 

As for the output today, yes it’s pants, and I have not yet been convinced of a very cold outlook. Really to have confidence in a very cold spell we need to have consistency across the big 3 - we are yet to have that and I still think we are looking at a chilly, messy picture with the chance of snow at times, with milder interludes, rather than a very cold outlook. 

I’m thinking December 2017 rather than February/March 2018. 

Still, due to the effects of the SSW the models are likely to be all over the place for a while still, in reality not much has changed from yesterday. Wouldn’t surprise me if upgrades were back tonight, and if not, tomorrow. 

People are making the mistake of becoming emotionally involved in the NWP, and making judgements too quickly based on daily op runs. Happens every year and we seemingly never learn (but we are getting better). 

Why do we even look at charts post t-144? - it is no better than looking in a crystal ball with Mystic Meg.

We think the ECM is more durable than GFS which to be honest has been flapping like a flounder out of water. But ECM itself is flapping around worse than usual - Brian's upside down chart above says it all.

Could be that Shannon Entropy is to blame, or the unpredictable consequences of the SSW, plus the GFS must be affected to some degree by the US shutdown.

The only thing we know for certain where UK cold is concerned - if it can go pear shaped, then it will.

I think when it comes to the ensembles, the ECM tends to be the better of the two (generally speaking) up to the 15 day mark IMO) but I agree with yourself and others that with regards anything beyond this, the ECMWF is no better than any other model. I have been keeping a very close eye on the EPS 46 day charts and outlooks since the Autumn and they are pretty much consistently useless beyond the two week mark.