I agree 100%. I did not want to muddy the waters further since people are having trouble handling the statistic so far.
What happens when you get rekt 5 games in a row by the same "unranked" top 10 premade team? You should lose NO ELO at all since they are really top 10 players. But the monkeys at Retardic have no concept of this so you may lose a ton of ELO. Does anyone know? Do we have proof and data to support it?
There are a number of things that make this whole process not a 50/50 system. We do not have the Relic server code. There could be a number of things programmed into it that harm certain players. We have no idea.
Lets look at this another way: How many game should a person lose before they get a win?
Should the server at a minimum try to give a person a better chance after forcing them to lose 5 games in a row? 10 games in a row?
You may lose tons of ELO, but you can also regain it back over the course of a week. It's a back and forth tug, but you reduce your chances of clawing it all back if you continually /l based on your teammates before a match begins or when things go bad.
It's really not going to be a 50/50 system in any form because of premades and also I don't think the ELO gain/loss averages out to 50/50 either, it's a very spikey system made worse by the relative poor quality of players at all but the top 50 in 4s.
I have gone on 12-15 game losing streaks and have gotten 10 game winning streaks, I've gone from rank 50 to rank 400 in an evening. I don't think asking how many games a person should lose before they get a win is a correct one given the context of competitive ranking even if it is a shitty one because the answer is always going to be: a player should lose as many games as it takes until they get a win. There is no room for pity mechanics here when you're talking about something that is binary in definition - you either win or lose. Now how much ELO you win or lose that can be up for debate.