That was a Fiorawillcarry post Some one who typically is less then constructive in their thread/posts.
5050 is a stubborn fellow. I showed I can climb to gold with 5% of the games he has but the dude still refuses to treat me with good faith. I really hope he starts trying to "game" the system like I did because it requires a dramatically higher MMR then promos to move up. His knowledge on Elo systems is terrible. Hell he cant even tell if some one is duoing or just on the same team as a player on op.gg. The guy just does not seem to care about learning the math or logic. Anything that does not fit his narrative is ignored. How many times did I ask him to respond to the linked 1000 game to gold account? Notice 0 replies because it directly contradicts his belief that if your "stuck" its impossible to climb not because of skill but because big bad rito wants to keep the man down.
Why? He doesn't care for anything that's not justification his for personal perception of skill vs reality.
Finds excuses, its To rig population distributions to different tiers and/or its to make the game more grindy. Yet.... rigging results breaks the measurement of skill, the same measurement that needs to be accurate... for them to be able to rig games. Cant use a forced WR to have WR represent an accurate skill. Cant use a Forced MMR value and still have MMR represent accurate skill.
Riot already has 100% control of population distribution as they control what their ranks mean in relation to the population curve. It is set off % because influx and outflux of players means no static value will be usable.
Grind? when the average player by definition requires a climb time to reach the top of the ladder in the relm of Eternity. Brilliant idea! Riot clearly is going to use a more complicated system that does not work mathematically to accomplish the end goal of infinity plus 1!
You had a reply to me that you did not post as a reply so I did not read it prior.
So I have about 100 games ranked and finally hit gold. We're in a similar boat it seems. So if I read you right, you're basically affirming the division chokepoint system that balances itself out at "enough games played". That does make it carrot on a stick for people who otherwise belong in the divison they're trying to get to. How do you define that?
Not exactly. The idea is that you want the weakest player in the game to have opportunity to beat the strongest in a "'fair" match. A game where one player can only ever lose is inherently unfair. The goal is to have ratings accurate enough to use this for your MM system to build fair games. So a best game on the weakest player can beat a bad or average game of the best player. Ideally you want a exact and equal skill for as many players as possible. Reality is you simple don't have infinite # of players any given second in any given que so you will not see this happen much. Hence using a Range for MM. Tighter range tends to be higher quality but as gameplays entirely on players a on paper fair and close game can go to trash and there's absolutely nothing MM can do to solve that.
Elo systems are built around class intervals and the standard deviation expectations. You need to break out of a reasonable deviation to prove you don't belong in a Class interval. Its more about the gap between wins and losses then any specific WR. 100% WR... for... 2 games is not a big deal. 80% out of 100 is pretty dramatic. WR is expected to sink closer to 50% as the expectation is when you represent the average player your going to be with peers so you see equal odds. 50% is not forced, its a naturally occurring value. Same way if your 150 pounds and you lined everybody up by weight, your going to be near folks closest your weight why its measured out correctly. Averages occur in nature, they exist even unmeasured. Elo systems are designed to measure them, its literally a Rating system.
Anyhow so the point is the folks who climb to a point they need to play their very best and also for the enemies to not play at their very best is when you get stuck and feel the most helpless because ya have to rely on a higher # of teammates to preform well or enemies to preform badly. Not great odds.
@transparancy. Its not a false grind but it can be disheartening. Something like 100-300 games for high system accuracy, less depending on if Riot added some League specific Bayesian inference. Imo the issue is riot gives ranks out very fast with out setting the expectation that at high uncertainty the system expects to be wrong and need to adjust if it was.
Large #s is a time proven method to remove RNG and luck as variables. That's why the house always wins.
Feel free to bug me if you have any interest in Elo and MMR. I can't always provide exact information on Riots thou I'm pretty good at research and tend to learn about things that I find nifty. I will be clear if its anecdotal or just in adherence to theory or if I am directly quoting riot.
End of the day, you have already climbed. Plat is a good deal above the starting MMR.
