The dog dug himself a small snow cave and I dug a quick snow pit which predictably revealed 40cm of very dense accumulated wind slab on top of a 4cm crust which was protecting about 50cm of extremely loose faceted snow. We were only on a 30 degree section of slope and the slab was pretty stable on that, although the column did fail slightly on the faceted snow as I was cutting it. Even though I guessed that the slab was strong enough to bridge our weight, with wind like that we decided to cancel the remainder of the ascent to the col.. Who knows what small pocket of weaker slab was building in one of the many gullies that we could easily have ventured into. After that we had to snow shoe back over a flattish section that was still full of rocks and patches of grass.
Our descent from there was unremarkable. I tried to find a protected couloir on eth way down but failed, leading us into a short rocky chute that still had a running creek under the snow. Luckily the old Roman road is very well sheltered from the sun and had snow all the way back to the trail head. This path gets quit icy lower down and it was certainly not what we had hiked 4 hours to ride.
Day 3: We took a day off. Our planned start time was again 7am and that is well before the day's avalanche bulletin is published. Considering the very high wind from the previous day I speculated that the official avalanche risk would be raised from 2 to 3 (out of 5) with the possibility in the bulletins text for warnings about localised wind slab. Plus we realised that the snow conditions were so bad that there was little point heading up to that area for the next 4 days running. It turned out that the risk level was not raised in our area of the Alps, the boarder between the Haute-Savoie and Savoie regions. But further to the south they had increased the risk rating and predicted local slab activity. I have since read that in that area a skier was rescued from an Avalanche.