It's simple. A win/loss LP mitigation system.
Here is my idea of how it would work:
1. The only way to gain LP is to win the game.
2A. Every game is a wager of 100LP: Meaning that the losing team loses 100LP divided up amongst the five players respective to their individual performance.
2B. Every game is a wager of 100LP: Meaning that the winning team gains 100LP divided up amongst the five players respective to their individual performance.
3A. Series games award LP that roll over into the next tear/division if a series is won or LP is deducted if a series is lost (Rollover LP): This abolishes the "vanishing LP" gains after winning a series yet starting at 0LP. Furthermore the "Rollover LP" if the series is won acts as the demotion buffer already given but now assists in climbing, think Kinect LP versus Potential LP. The Lost series rollover LP deduction is already in existence--which I would like to mention that this has created a built-in negative LP trend, think "the house always wins".
3B. Series "Rollover LP" are subject to 2A and 2B individual performance laws: This determines the amount of Rollover LP gained/deducted based upon the outcome of the series.
Abolish MMR: MMR is longer needed in this system as it has a built-in MMR based on individual player performance on a game by game basis and the player's average play would form a built-in MMR average over X number of games more than one.
1) This system creates a faster convergence for a player's average played "skill" level. If a player hard carries a game they gain more LP thus converging to a player base closer to the individual's average played skill level.
2) If a player performs really bad and is carried they gain less LP thus keeping them from climbing into a higher average player base skill level keeping the carried player from potentially jeopardizing other players chances of gaining LP because they were carried into that higher player skill level and are not performing at that average skill level.
3) This system represents an individuals average played skill level better than the current one.
1) Smurf lanes would be "punished" harder by naturally performing worse due to playing against a smurf and receiving a larger share of the lost LP or gaining less LP based upon the Laws of 2A and 2B.
2) Internet going out would affect a players LP loss more so with the new system than the current one due to the 2A and 2B. (However, that is only for a legitimate technical difficulty. A Rage quit player would, as they should, receive a greater/lesser share of the 100LP wager based upon the outcome of the game.)
The BIg Question: What determines individual performance? Simple go to your match history click on any game and then go to advanced details. With that data alone an equation could be formulated to calculate a player's share or gain/loss of LP based upon the relevant previously stated laws.
Closing thoughts: The current system represents the average player base of a player's current rank more so than the individual player due to the negative LP trend inherent in the current ladder system. The negative LP trend is caused by vanishing LP in won series and losing a large amount of LP in lost games regardless of playing well caused by the lack of performance of other players. This system would only be implemented in Solo/Dou Q for obvious reasons. I would like this to open a conversation of ideas of improvement for the current ranked ladder. Please share your ideas, add to pros cons, or venture with me to further develop the concept.