You mean if the match up is bad? Your odds are never so low that resigning is worth it before the game starts I don't think. Unless you're freeze mage vs warrior.
Yeah. I see it often enough that I figured it was theorycrafted out.
Ranked play is attempting to win your way up the ladder, and your win/loss record actually doesn't matter whatsoever. So it comes down to using your time as efficiently as possible to get more wins than losses.
For example, in the time it takes me to lose a 27 turn game to a Control Warrior, I could have gone 4-1 or 3-2 against a number of other opponents. Even if I only get two wins and no losses in during the time it would have taken me to lose the game I resigned, I'm ahead on stars.
For me, it seems like the decision tree looks something like:
#1 - Have I won two or more games previous to this one? If yes, play the game. If no, go to #2.
#2 - Is your opponent's class, regardless of deck type, one that will result in a game of >X turns more than X% of the time? If no, play the game. If yes, go to #3.
#3 - Is the amount of time it takes to play >X turns greater than playing X other games?
And so on...I don't know all the right questions yet, or the right numbers to fill in the Xs, so I was hoping the work had been done.
There may be other variables (i.e., is your opponent golden? etc.) to take into account for sure. And this is all pre-legend, so I'm sure there's even more though to be put into it depending on specific rank. Then again, if you are Legend, the stars thing doesn't matter and the level of competition is invariably better, so there's even more to compute.
Basically, there's an equation that weighs the time you have to spend to possibly/probably lose. So for example, if I lose 60% of all games against Warriors, and 80% of all those losses take 10 minutes or more, it's more cost-effective to just resign all games against warriors before wasting the time in favor or seeking better matchups (numbers still made up).