If we are to take subsequent studies at face value, then it's true the Tories would have actually done slightly better in 2015. I personally think that doesn't butterfly 2017, as having half a dozen more seats still isn't going to insulate May from the brexit related rebellions she likely had in mind when calling the election.
And this is where things get interesting. According to ERS, Labour would
have done two dozen seats better in 2017, with the Tories losing 13 from OTL. Interestingly, the SNP would have lost exactly half of the 54 seats they would have gotten in 2015. What would have happened from there has already been litigated in other 2017 GE threads, but it's clear that a Conservative government would not have been viable.
Of course, voting behaviour would change too. My guess is that we'd see more votes going to parties like the Greens and UKIP/BP as first choices. They wouldn't win more seats because of that, but it could have an important effects for the way they are covered, and possibly in terms of Short Money too. It might also prevent BP standing down for the Tories in a 2019-type scenario, if that ever arises.