And yet isn't HOT's adding something like this to ranked? So it's a different story maybe depending on the TYPE of game. 3rd person vs 1st person, laning-based vs non-laning-based game. Also, there's no itemization and leveling up abilities in overwatch, but there is in league and hots. If it can be done PROPERLY, I am entirely in favor of performance-based Ip gains or losses. If I'm getting bot lanes that go 1/7/0 in laning phase, shove blindly without warding, ignore mia's and danger pings, and FORCE fights vs enemies they fed (as though they stand a better chance of winning now that the enemy has 1500 gold more in items and they couldn't win when it was even in gold); I want to lose less Ip for this if I perform above average. The problem is, much of this can be immeasurable by statistics. Kda can be a horrible indicator. Gold farmed - well, we've seen in the pro scene some of the best adc's don't actually have the highest gpm, but when you look at their opponent's gpm, it is half of what is expected from that opponent because that best adc zones the opponent off farm so well.
Similarly, you might have low kill participation, but be winning a difficult matchup while being camped by the enemy jg or constantly roamed on by the enemy mid and bot laners or top laner. So it LOOKS like you did little to help your team, but in reality, you absorbed a TON of pressure, remained relevant, kept up map pressure as much as possible, and still won 2v2's or 1v1's whenever not being ganked. Or maybe, you got your opponent's tower really low w/o any help from your team, but then due to need for lane rotations an inopportune fight causing you to have to leave, you miss out on killing the tower and the gold it gives; then it just goes into stats as an accomplishment by minions or your teammates while it doesn't look half as amazing in your stat line even though you did most of the work. Conversely, you could feed all of laning phase, get carried by your teammates mid game, be relevant in terms of damage by late game, carry ONE fight all game, and win. Then, if you'd "ks'd" well in those mid game fights to become relevant, you have an amazing kda; meanwhile, your teammate who had to carry those midgame fights and die every time so your team would come out ahead has a mediocre kda by game's end even though he/she did much more work than you. Thus, your Ip gain might be inflated while his/hers might be lower than it should be based on actual performance. Stuff like that is the only problem I can see with making Ip gain/loss performance-based.
is known for having smoother aa animation, which impacts CSing.
Steel Legion 
has harder-to-see visuals, which makes opposing players react to skillshots slightly slower than usual. The 0.01sec can be the difference between landing the snare on enemy ADC 50min into the game and destroying their nexus, or not landing the snare and losing the teamfight&nexus.
Dragonwing 
's E particles are extremely misleading for enemy players; it's hard to tell exactly how much the AoE covers.
