Look theres no need for a forced 50% rate ever. That system is unrealistically difficult to build. Your asking for A predictive algorithm that inherently denys the core functionality of Elo systems. In Elo/MMR you use the results of a match to build the ladder. You adjust players at the end of a game. If you force 50% because your systems proactively trying to manipulate the interaction of a game you destroy the accuracy of if a win or loss is from player skill. Thus it wouldn't be an Elo system, then we wouldn't see players at the top and bottom ends of ranked ladder sorted out by skill.
Instead allow to me to explain that its really easy to abuse a MMR system to accomplish the same end goals you would EVER want to accomplish with forced 50%. Especially with the additional buffers in Riots League system.
**Don't want majority of folks high on the ladder? **Make the starting point for the game low. Starting points going to have roughly 50% of the games playerbase below it in Elo biased systems. So the starting MMR for a new account can just be set to bronze or Iron.
Don't want people to climb fast?
Have a Low starting K factor, Have it lower rapidly over the ladder.
Want to" hard stuck" players into low mmr?
Change the MMR adjustment algorithm. Make it so accounts lose more MMR then they gain. Prevents any one who cant keep a high enough win rate from climbing.
Want to make people dramatically rubber band around the ladder? Make it so that a win streak or loss streak feels really dramatic?
Set a High K factor. Low slowdown.
You can further increase the rubber banding by setting points by MMR where your adjustment formula changes to favor wins > loses for lower ladder MMR gains. And high Sees more MMR drops on losses then wins.
That would get people to climb rapidly off a win streak. Regardless of how good they actually are their moving rapidly thru divisions, but at a point It stops and flips.
This makes average players bouncy, never at the bottom, never at the top, never able to climb to far from their skill. Yet constantly in dramatic motion.
It also makes matches a lot more uneven for the majority of the players.
Want to make a Time sink?
Don't add Speed up modifications to shift folks around ladder faster. Oh, and again set a low k factor. A very simplistic MMR system for 5v5 team play could require several hundred games to get reasonably accurate for any 1 account.
Riots last official word was they expected their system to require about 150 to 300 games depending on RNG of luck w teammates.
Want to sell more skins!
… I have no idea. Honestly 0. I see no relation between matchmaking systems and skin sales. If its a side effect of anything I said above then just do the above to sell more skins as a side effect. If you think its related to something else, that's directly tied to matchmakeing, list it and I will try to address it.
Look. Elo systems can be adapted for team play. Im happy to explain how if you need help. They do work, like their entirely usable for making a viable matchmaker and ranked ladder.
They don't work if you try to proactively rig matches. They don't work if you remove skill as a factor. They don't work if you force 50%.
I have never see a counter argument that's got solid logic yet, Like I just dont understand it. How do people say in good faith that its easier to predict the future then it is to just make a suboptimal elo system?
So please, anybody who believes strongly in forced 50%. I invite you, Go ahead, if you have something you think Riot would gain with forced 50% rates. Tell me what you think it is, long as your realistic in your expectations I am pretty confident I can accomplish that same end goal with out even needing to predict the future, but by manipulating reactive statistical systems.
*
