There was a bit of a debate on twitter last night about whether the Tories were more likely to vote with the SNP. So I whipped up a python script to parse the data and do some quick and dirty analysis.
2011 is missing from this analysis as http://www.publicwhip.org.uk/ is kaput presumed in need of love.
Methodology: Count all the votes. If more than 10 government MSPs and 10 Tory MSPS and less than 5 of the main opposition MSPs voted the same way, count it as Tory collusion. This eliminates motherhood and rhubarb pie votes where everyone agrees and times when an MSPs finger pressed the wrong button or something.
1999 Tories voted with govt 23 times of 103 or 22%
2000 Tories voted with govt 88 times of 277 or 31%
2001 Tories voted with govt 74 times of 329 or 22%
2002 Tories voted with govt 117 times of 369 or 31%
2003 Tories voted with govt 67 times of 374 or 17%
2004 Tories voted with govt 115 times of 383 or 30%
2005 Tories voted with govt 68 times of 334 or 20%
2006 Tories voted with govt 110 times of 379 or 29%
2007 Tories voted with govt 77 times of 256 or 30%
2008 Tories voted with govt 84 times of 217 or 38%
2009 Tories voted with govt 77 times of 212 or 36%
2010 Tories voted with govt 28 times of 71 or 39%
As a quick and dirty metric, I think this shows that the Tories have found much more common ground with this SNP government than they did with either of the previous two Labour-LibDem governments.
ETA: this is just a bit of fun, the SNP these days are a centre-left social democratic party, though they have been dragged to the right a bit by the necessity of working with the Tories. Life in Scotland would be a lot better if my tribe and their tribe could put aside their differences after May and work constructively together against the vicious Tory-LibDem government in Westminster.
 You can find the source code and all the data for this at https://firstname.lastname@example.org/aidan/vote-parser.git patches welcome