New ranking system to stay or?

[Jules]

@Jules no offense mate, but this is kinda the In-House league section. So this system only exist withing the closed In-House League games.

I was just offering my thoughts on Elo vs win rate, as someone who has had experience with distinguishing those things and explaining the math behind them. Not trying to tell you guys what to do!

[Dong]

np's

[Krayyzie]

Just a quick side note,
The reason I posted a few days ago i was against keeping this system with random teams and change elo to be 17-13 and 20-10 and 23-7 and so on, is because with random teams ,a majority of games will be 20-10 or more, and in an inhouse league with players that needs some nesssesary skills to get into, noone will ever win 65-70%, which means all games will benefit the players with lower elo. With random teams, keeping taht system would screw up the ranks even more.

[HealByColor]

A lot of the game is luck. Especially if you get a bunch of people around the same skill level. Sending is probably one of the biggest factors of the game which can be controlled by the players. Archers are good but if you can bust them up early pretty much no different then the average yolo tower. I would suggest watching kill rate and etc to send better.

[Jubadus]

A lot of the game is luck. Especially if you get a bunch of people around the same skill level. Sending is probably one of the biggest factors of the game which can be controlled by the players using mh. Archers are good but if you can bust them up early pretty much no different then the average yolo tower. I would suggest watching kill rate and etc to send better.

[RadiantCrystal]

Please stay on topic people

[FadingSuns]

Random teams is one of the biggest funs i have found so far in this league. Reasons:

1) I want to play with all players
2) Beat a Stacked team is kinda challenging

I dont really see any conflict on having to make calls, control King or plan the strat. Its part of the game and everyone should know how to do it. Another thing is that i always prefer to let others do calls or control King (depending on who is on my team). But it should be a MUST for all to review the call made by the "leader" and state anything if you disagree with it or think that another plan would fit better.

In any case, we should not be afraid to state our thoughs regarding what i just said. Also, i can see that people like 1007 is kinda tired of making calls everysingle game, taking the lead of the calls should be a thing that we all can do.

I agree with this system, i agree with random teams and i agree with the 15/15 Exchange.

Getting back to the case about having lower elo with 47% win rate than other with 30% winrate should not mean much srsly. I dont care whats the elo on King, i know he is good player and that is all that matters to me. Elo or winrate can say many things, but i am enought grown to know if a player is good or not based on experience playing together or VS rahter than elo or %win.

When we try to evaluate someone individually on a team game those should be part of the analisys but cannot ever be all. As i said, game experiences is (for me) the most usefull way to evaluate the skill of a player. Dont care that much about elo GUYS! :=)

[eldryan]

I think w/l rate isn't the end all of skill level. It's easy to go on streaks then go inactive, sustained success means more even with a worse w/l in many cases. 400/225 is better than 200/100 is, generally. OccupyWstreet went 70/40 one season, but it wasn't the best season.

My preference was Elo based off 4/26 etc. If you deserve your Elo, stop stacking more than half games.

[RadiantCrystal]

I would say random teams are good, though the elo should work normally so it offsets the potential "skill difference" It's not impossible to beat a stacked team, but the probability of winning vs a stacked team is rather low compared to a balanced team so the elo should act as a "reward"

[KinG23]

^ My thoughts exactly, even though I still prefer the old system.

[Jamo]

I'm with RadiantCrystal! Random teams are fine for me, I actually like that!
But I see no point in staying with 15/15 elo right now. It's my first season and I have around 800 elo atm (sad ) and usually I get really motivated when I am in the lower skilled team and have to face four really good players. Then I think, "wait, this would be a 27/3 elo game and if I win, I just get the half of it and if I lose I lose so much more only to give the high elo players even more elo". That's completely demotivating and imo unfair. The difference between low and high elo will be higher than before, that's something not to really care about, because it's relative. But with a game to game view, I really would like to get or lose the elo I "deserve".

[n1ll3]

I dont really care about the elo system..

Only 2 points are important for me:

1. Random teams are fucking nice (i want to play with different people).

2. The maximium elo change should be not more than 20/10 in a league with only skilled players( more or less ).

[KinG23]

Any updates on what the plan going forward is with match making? Maybe make a poll? General consensus seems to be go back to the old system, or keep random teams using actual ELO with a maximum of 20/10 per game.

[pewpew lasergun]

after 2-3 months of play i think we can all agree that
this system punishes the low bracket players
and helps the mid-top bracket.

This season will have a new top record and low record.

I am not sure which will be in force next season, but we can even alternate maybe. both random and elo games i think are good. a poll wont hurt i guess.

[supersexyy]

after 2-3 months of play i think we can all agree that
this system punishes the low bracket players
and helps the mid-top bracket.

Wasn't that the point of the system? Obviously this system is a less accurate portrayal of your elo than the previous system.

The only real benefit of the new system is the randomness of the teams which allows you to play with different players. This can be achieved by randomising teams and retaining the elo scaling of the previous system. This will result in greater differences in elo which can be counteracted by scaling the elo formula to be less harsh.

Tldr Having a flat +-15 mmr is pointless.