My point is that his E is not supposed to be a difficulty of his kit. It's not supposed to be way harder to pull than push with this skill. It's obviously made to be a easy-to-land skill (no cast time, short but large...)
Riot wants that the complexity of the game to be about choices, not about useless difficulty. (I'm merely quoting something a Red said a long time ago on the forums, don't remember who and what exactly, but eh, just to say I'm not inventing that out of nowhere :3)
And yeah, I think that if you HAVE to aim a skillshot elsewhere than where you want it to go, then it's useless difficulty and not complexity...
My intent is that it is "harder" to pull than to push. This is not because I like "useless difficulty", but rather because I had to make a choice as to which mode would more difficult. The reason I kept the ability with the targeting the way it currently is is as follows:
*In an "oh shoot" moment, you want the most intuitive outcome to be one that is less likely to result in you dying and more likely to save your life. Freaking out and pressing E should not draw the enemy in; pulling in should be the more calculated choice.
*Thresh's E combos with his Q in two ways -- you can EQ to set up a Q pull with less chance for the enemy to dodge or you can QE to extend the effect of the pull. EQ needs to be executed very quickly (button press within 0.5 seconds of the other) and Q is narrower than E. QE actually requires that you delay the execution of the E (~1.5 second delay) and the enemy's position is locked as you have them Q stunned. So these combos dictated making the QE combo the one that required the most mouse movement.
*Players tend to, rightly, think of themselves as the locus of their abilities. Having the vector -- the direction -- in which the push occurs be the direction from Thresh to his cursor felt a lot more natural, though admittedly not perfectly natural. It was very, very difficult to execute the "cast behind to knock away" version of the spell with targeting reversed.