I really don’t understand why the system is designed like this.
For example,
A Guy’s ELO is 1200, rank is Platium I. (A guy was 100% win rate in the beginning 9 matches)
B Guy’s ELO is 1050, rank is Diamond I.
A guy has to face good players with ranking Diamond II ~ Conqueror I and they are with ELO 1250~1350. No matter how many games A guy played, all the opponents are at least Diamond I and with high win rate. Which makes A guy really difficult to win and even drop to Gold III.
B guy can easily reach from Diamond I to Conqueror because of his low ELO. His last opponent is even a silver II with 6 games and 50% win rate. WTH?
Is this system designed to ask players to try to lose their begin 10 games to get a super low ELO for the ladder?
Has the designer ever tried to simulate the system???
The rank system is so broken.
And they didn’t even fix the unintentional points deducting everyday like 10 days already.
I wonder what they are doing at work.
even though the leaderboard server sends the rank points disguised as elo, i think the formula itself is different compared to the actual unranked leaderboard where they use the standard elo formula.
I’d still like to know how to see these two pieces of information for each player. As the “elo” in that json is the same number as the rank points, I suspect they’re the same piece of information with the elo being an incorrect name.
yes, they use hidden elo on the backend for matchmaking, but send rank points disguised as elo to the frontend. i cant say much about the good/bad of rank points, but atleast they should disclose the hidden elo, so people have some idea what to expect.
I think the OP must know of some way to see both pieces of information, as they seem to be talking about real players rather than hypothetical examples (it’s the reference to the last opponent’s details that makes me think that).
Thanks for the info. This clarifies alot of things.
I thought that Elo is the overall skill of a person like in chess. But here you have elo for every different mode you play so to measure your skill.
I would have kept the same elo for 1v1 Quickmatch and Ranked because the skill is the same.
The problem here is that it’s confusing to have one number for matchmaking purposes (elo) and another (Rating) for standings.
We also use ELO to determine how many Rank Points you gain or lose.
So Rank Points are just a “graphic” visualisation of the hidden elo.
All this shows that you can get lucky at those 5 placement matches as a low elo player and get placed way above your skill. Might be wrong but thats how i see it.
And like the Op said if you are a low elo player who was lucky on placements you have an easy life climbing.
And this is because you are matched based on elo not Rank points (as i thought initially)
Does this mean it’ll be super hard to get high rank in both season 1 and season 2? Since I’ll be starting with a high elo and 50% win rate approx, so will have to just grind the +2.5 points average per game. That’s a lot of games to Conq 3.
It’s not completely clear exactly how it works, that’s why I’d like to be able to see both elo and rank points for everyone on the leaderboard to get a better idea.
They say
“your ELO in the Ranked 1v1 queue will persist and will play a role in determining your starting Rank after your 5 placement matches are completed in future Ranked Seasons”
It isn’t clear exactly how the persistent ELO will relate to the starting rank points, but it seems safe to assume that higher ELO players will start at higher rank points than lower ELO players, even if it doesn’t go so far as to set rank points to be equal to ELO initially.
ELO is used to track your skill in each matchmaking queue. This means that you have a unique ELO for Quick Match 1v1, Quick Match 2v2, Ranked 1v1, etc.
This corelated with the information from Quick Match :
You’ll need to complete 10 qualifying matches for your initial ELO calculation and placement on the Leaderboard
It’s safe to asume that at the start of season 1 the 5 placement matches not only determined the Rank Points but also your initial unique Ranked 1v1 Elo. So everybody had the same elo before placement matches.
This unless they used your previous Quick match 1v1 Elo. But i don’t see this info anywhere.
This kind of explains the problems season 1 has.
your ELO in the Ranked 1v1 queue will persist and will play a role in determining your starting Rank after your 5 placement matches are completed in future Ranked Seasons.
This can potentially solve the problems season 1 has. But i still believe some lucky low elo will be still placed higher.
This is why they need to be more explicit on how the system works. And be transparent with the Elo formula.
Especially because they will introduce Team Games in the future Ranked Seasons.
Without the ability to see ELO, you can’t assess the fairness of all this. Gold and platinum players can be the same ELO as a bronze player, but just with more games played, as rank points increase with games played, even with a 50% win rate that won’t change your ELO (I’m assuming a 50% win rate won’t change your ELO, but without the ability to see people’s ELO we can’t be 100% certain of how any of this works).
Look at the peak time of the game, it is more likely that the conquerors already have a game and you will have players from diamond down, some conquers like leenock or 1puppypaw, they do not play in that peak time
You can get some insight into the relative ELO of players from the number of rank points gained or lost. I was just looking at the game history of a player who has around 1000 rank points. They lost against two different players, both in the 750-800 rank points range, but they lost 26 rank points for one, and 17 for the other. Clearly the one they lost 17 points for losing to has a higher ELO than the one they lost 26 rank points for losing to.
So I look at those two players, and the one they lost 26 points against has played 200 games with a 50% win rate and is gold rank. The one they lost 17 points against has played 27 games with a 63% win rate and is silver rank. So we can see quite a big difference between ELO and rank there, as the silver rank player clearly has a higher ELO than the gold rank player. So why did this happen? Looking at the silver player’s match history, they lost 8 of their first 9 games, then didn’t play for many weeks, then started playing again and have won almost every game they’ve played since then. I’d guess they played on a different account in between. I’d say it’s not unreasonable for the system to struggle to cope with a player who exhibited a huge change in skill from one game to the next, it can reasonably assume that in general a player’s skill improves gradually over time.
