Remember: Muster for Battle had the highest win rate when played on curve out of any card in the history of Hearthstone according to Blizzard.
Man that has me thinking about the enormous amount of data they can collect and the huge number of options they have for analyzing it. You could, for example, find out what the best "perfect curve" combo of cards is for the first three turns (how high does this winrate go I wonder?). Or you could see how the winrate of certain cards changes based on playing it before curve (through cost reduction or innervate type cards) and after curve (didn't draw until after that turn). So you could see whether Rag's win-rate has any statistically significant changes based on whether you play him on turn 6/7/8/9/10 for example.
In particular, I am very curious about the inherent variability of deck win rates with respect to curving out. That is, how wide is the discrepancy between the win-rates when you perfectly curve out and when you don't. You could end up with a scenario where a deck is balanced in the sense that all its various draw orders average a 50% win rate when summed together, but the win rates of specific draw orders may be wildly disproportionate.
As they mentioned in one of the Overwatch Dev videos/posts, players have a strong psychological preference towards winning by huge margins even if it means the match wasn't competitive. We think we want competitive balance, but players hate to lose close games and they don't enjoy winning a close game as much as they enjoy winning in a blowout.
This same problem may be reflected in Hearthstone through the statistical chances of your deck aligning more or less towards your 'perfect curve' (and its presumably much higher winrate relative to the 'average' draw order). Blizzard could likely calculate at the very beginning of the game what your chance of winning or losing the match was based on the deck orders of the two players. They could even go a step beyond that and measure the meta-data of competitiveness; how often was the win-rates of the two players near 50% and how often was it lopsided towards one player over the other, say 70-30.
Basically what I want to know is this: Is the average game of hearthstone more likely to be a competitive match whose winner is determined by individual player choices or is it more likely to be a blowout determined primarily by the draw order of cards? And is the answer to that question different in other card games or is it just a problem inherent to the genre?