Winrate gravitating towards 50% is mathematically correct, sorry!
Lets use this analogy: When you are a new player you will lose the majority of your games. This is synonymous to digging a hole. The lower your winrate, the lower the hole and vice versa. However regardless your hole is always going to be consistent in that it will only get wider. The wider the hole, the harder it is to change the hole's depth. The elo system comes with a metaphorical "shovel modifier", the shovel being your MMR- more specifically your teammates and your enemies which changes proportionally to your winrate. Your shovel makes it either easier or harder to change your hole's depth, but it does not take into account your hole's width.
What this means is if you start out as a new player, your shovel modifier is going to be average (50%). Your holes depth is going to become deeper and deeper, receiving worse and worse shovels at certain bad milestones all while your hole's width becomes indefinitely wider i.e. harder and harder to reverse these changes.
**What you guys who swear up and down that a 50% winrate is a conspiracy ** fail to understand is it's going to show up after many hundreds and hundreds of games because the more games you play the less deviation your mmr is ALLOWED to have mathematically. This means that MATHEMATICALLY your winrate will gravitate towards 50% REGARDLESS of who you are as long as you are not **A: In the upper extreme of players ** **B: In the lower extreme of players **
You can see the exact same pattern in random number generation. You can roll a dye 500 times, at the beginning of this dice rolls you will notice the numbers are skewed for the 1-2, 3-4, or 5-6 positions. Going back to the analogy I used, as roll the dye more times you are widening the hole- that is always going to be consistant. The numbers that come up after 500 dye rolls ALWAYS end up gradually progressing to a 33% chance of rolling one of three outcomes.
In league of legends, the random modifiers are your TEAMMATES and your ENEMIES. That's two variables completely independent of each other, one of these two will win. 50% chance.
But with league its different because unlike dice rolls, your two independent variables change STRONGLY depending on the beginning of your account's progression.
You can play under 500 games and have above a 50% win rate all day long because when you go on win streaks it counts THAT much more than say if you play 5000 games. When you get win streaks that count that much more, your MMR INCREASES proportionally to your winrate. Thus you get a net gain of both better teammates AND better enemies, which means you are far less likely to lose because of silver 5s who you still see because that's where you started when the account was made and where you started your climb. This is the drawback of an elo system is HEAVILY WEIGHTED from GREATEST to smallest on 1.) How many games you win or lose within your first few games which effectively determines your average game's skill level(the shovel modifier). 2.) your total games played which is your "hole's width", 3.) your total winrate for ALL GAMES(the function of your hole's width and hole's elevation/depth), and FINALLY 4.) Your total skill progression
League of Legends is NOT a wild exception to this matchmaking flaw either. MANY games use a similar system that calculates player skill like this, you will notice in these games that THIS ISSUE IS NOT UNCOMMON.
tl;dr:
Mathematics prove that through trying to determine someones skill level, which is abstract- with initially specific perimeters that then determines RANDOM NUMBERS(your teammates and enemies skill difference NOT including yourself), that this persons skill level will ALWAYS come out to being "average" after a certain point because of the dampening effect that naturally occurs with random numbers that are played hundreds and hundreds of times. HOWEVER, this is not the case if the initial perimeters are either extremely high or extremely low because the higher your teammates mmr the more likely you will not lose to dumb mistakes that random teammates are known to have thus your skill expression is better rewarded. While the inverse is true if they are low than your overall skill expression has a much lower impact on the outcome of the entire game.
