not widely beyond the usual margins of error, but as I've pointed out before, the D'Hondt system, with its added uncertainty of how regional list votes play out, makes elections particularly hard to predict, and a percentage point or two either way can lead to vastly different outcomes.
All that said, here's the latest poll by YouGov for The Times which may prove prescient or (more likely as I see things at the moment) an outlier:
Scottish election: SNP and Greens set for historic election success
...
The YouGov research for the Times forecasts a four-seat majority for Nicola Sturgeon’s party.
The pro-independence Greens, meanwhile, are on track to more than double their return, with 13 MSPs. That would put them just four seats behind Labour, who are predicted to lose out to the Tories in the battle for second place.
Polling expert Professor Sir John Curtice used the survey results to draw up seat predictions for Thursday's vote.
According to his analysis, the SNP would win 68 seats, while Conservatives win 26, five fewer than in 2016.
Labour would win 17, down seven, the LibDems four, down one, and Alex Salmond’s Alba Party would return one MSP.
https://archive.is/BK2ZY#selection-1515.3-1523.36
Other polls have made the election much more of a squeaker for the SNP, varying between just missing an overall majority (in practice almost certain to be topped up by the Greens to form the next government), and a bare one- or two-seat majority.
The Alba Party have varied between 2 and 6% of the nationwide vote. Cagey predictions on that basis (they're only running as a regional list party, so it's the percentage in an individual region that would matter) have veered between no seats at all and maybe 4 or so. Curtice has (perhaps overgenerously, but he's the seasoned pro) awarded them 1 seat with 3% of the vote
= new reply since forum marked as read