Hi Gav
I have just watched the 2019/2020 forecast all the way through and thanks for all the hard work you have done to get this out and for all the seasonal parameter and model estimates over the past few weeks !
I liked the way you chose the analogues and used the reanalysis app to get to an "average" set of height anomalies to predict this winter. Of course you do have to start from the premise that weather has this sort of "memory" - in my opinion there is some validity in it but it can break down without warning sometimes as one regime ends and a new one starts (but not in a predictable way)
I was so taken with the reanalysis tool that I tried it for myself in a different way to you
I went to the CET database and looked up the 7 coldest amd 7 mildest winters between 1836 and 2015 (the date range of the reanalysis V3) and then compared them with your and Terry's selections of years as analogues to 2019/20
The results were (clickable to enlarge)

As you can see both your analogues are very much closer to the cold winter pattern than the mild and I would be tempted to say you and Terry have both been a bit over cautiious in your interpretations
The only (big) caviat is the SST as you already pointed out
In particular the SST warmer patch in the N Pacific between Alaska and Russia is IMO a disaster for cold winters exactly as you described. It anchors the Jet Stream onto a NW/SE trajectory over N America and forces a very zonal flow across the Atlantic with intensified SW/NE storm tracks - ugh let's hope not
Otherwise I might have been tempted to "go for it" this winter