I was asked today whether it would be possible, even very speculatively, to estimate what the electoral cost to the SNP might be of deciding not to pursue a full-blown Home Rule Alliance with the Greens, SSP and other non-party groups. I don't think it is possible, because those of us who supported an alliance weren't hoping to merely "tack on" the small Green vote to the SNP tally, in order to get us over the line in a few constituencies. That was part of the idea, to be sure, but it wasn't the primary motivation. (A bigger red herring was the claim that an alliance could only be of any value if it could be demonstrated that there is at least one constituency which the Greens or SSP are better-placed to win than the SNP. There is of course no such constituency, but that simply isn't the point.)
The real hope was that an alliance would be greater than the sum of its parts, as the SDP-Liberal Alliance was in the 1980s, and indeed as the Yes movement was last year. We thought it might help to attract traditional Labour voters who still nurse hang-ups about the SNP, and perhaps also some of the semi-mythical "missing million" who wouldn't otherwise turn out to vote. By definition, the extent to which any of that would have happened will always remain unquantifiable.
In any case, the debate over a potential alliance mostly took place before the SNP surge in the polls became fully established. It's possible that we overestimated the hang-ups that Labour people have (and Nicola Sturgeon's leadership may be helping on that score as well). It's also conceivable that the SNP's "brand identity" is so strong that the party is actually faring better on its own than a new political force with an unfamiliar name would have done.
Basically, we'll never know for sure.
* * *
Today's Scottish subsample from YouGov shows an SNP lead of 43% to 25%. Friday's result was very similar. So the little flurry of narrower gaps that we saw a few days ago does look like a blip.