That rank system blog post wasn't very good or helpful

I really don’t understand a system where hidden stats are what dictate ranked matches, not the RANKS themselves.

A much simpler system should be in place in my opinion:
10 placement games. How many of them you win dictates which rank you get placed in. Win 0-6 you get put somewhere between Bronze 1 and Bronze 5, 7-9 of them gets you between Silver 1 and Silver 5. Winning all 10 puts you in Gold 1. You should not be able to place any higher than that.

Then literally just winning a game boosts you a certain amount of rank points, losing sets you back a certain amount of points. Take performance out of the equation.

Matchmaking should be locked to anyone in the same skill rank, one rank above, or one rank below.

Hidden stats may theoretically create fairer matches on paper, but for rank progression and overall ranked experience it makes the system feel unfair, confusing and punishing.

Good news about the Champ rank returning at some point though.

Now we also really need more ranked playlists:
6 player FFA
2v2 Slayer
4v4 Attrition (Either 1 or 2 second revive timers. Win 3 rounds to win the game)
4v4 Arena (Oddball, Capture the Flag, Strongholds, King of the Hill)
4v4 Slayer
Seasonally (Or even monthly) rotating ranked playlist: FFA Snipers, 4v4 Team Snipers, 1v1 Slayer, etc.

4 Likes

The blog post was great. Hopefully the will not is up with more info at a later date. So many questions still.

They did a good job of explaining why they have the two ranks. But the important bit is that most of the time they are pretty much the same.

This would take forever to rank the population.

And matchmaking would break without accurate ranks.

Onyx level players smashing Bronze would not be good for player retention.

Agree.

Your MMR isn’t really hidden. It’s pretty much there in your CSR.

And it’s not theoretical… The matches are much fairer with a more accurate ranking system.

I agree there is no satisfying progression. We really need an XP rank for that.

ELO type systems are very good at what they do. They can be confusing (mathematically) but they are not unfair or punishing. They just tell it (your rank) as it is.

It will be interesting how they do it.

The top 200 in itself is a flawed concept. There is just too much separation of the population.

Maybe on a regional basis?

Or a percentage of the population?

I would like them to divide Onyx into 10 tiers and name them after the to 10 ranks in Halo 2.

I think the most important part was them acknowledging some of the things people have been talking about for quite a while. They also confirmed some of the things we knew from experience already or had assumed.

Now whether or not we like some of those things is of course a different discussion. But I like that they came out and talked about how things work.

Yep. Good start… but hopefully they continue from here.

We need better post game explanation of why you only got x/15 to your CSR. It doesn’t have to be super detailed… but people will be less upset if they can see why they got 2/15 and their team mate 10/15.

eg. You got 8/15; win+, rank difference++, performance+, MMR drift-

We need a better appreciation of the weightings and how they are applied in Objective vs Slayer.

In the blog they specifically mentioned a case where MMR goes up despite losing. Is this just in placement? What personal performance weightings are we talking about (kill rates would make the most sense). How is this adjusted for objective games?

Are they happy with the way CSR drifts up and then crashes down on a loss. I know this is how it works when you positively bias it (giving CSR for eg ferry win) - but it frustrates the Hell out of people when they feel punished.

Losing CSR when someone quits?

Fingers crossed we get more!

1 Like

I was pleasantly surprised by the post and found most of the information very helpful.
However I missed one major topic: the way the system favours K/D as a CSR tool while not taking into account objective play. I actually prefer objective game modes in ranked over slayer but it sure feels awful knowing you sabotage your CSR when picking up the ball.

I think it’s K/minute more than K/D.

But yes, we need to know how it’s taken into account with objective games.

I agree that it shouldn’t be over-complicated, but at the same time, rank designs should consider all possible variables, which are missed if it’s too simplistic.

I don’t think 10 games can determine fair matches, especially if you just get wrecked cause you’re A) drunk or B) matched vs Onyx players. I would not be against them just removing the entire concept. Nobody wants to go straight into over-competitive matches where they die from all sorts of BS. Ranked needs players to have that period in their ranks where they feel like they’re worth a damn in the game, so they become slightly addicted to performing - similar to how you suggested that players start at lower ranks.

If K and D are the parameters built into the system… All they need to do is reward objective play with a K.

eg. Every 20 seconds holding the odd ball earns you an internal kill. Every 3 flag returns. Every 5 holds. etc.

I’m sure they could work out from game data what the appropriate numbers are.

Or just allocate some K on the basis of your overall objective score.

The important thing is that you know you are going to be rewarded for the team things. And you don’t have to run off and be toxic about your K/D.

2 Likes

It can.

Sure, it needs closer to 50 to be super sure. But it can ball park most people in a handful of games.

That’s on you. If you’re getting hammered and the then going into placement games I suggest that there is a deeper problem. Medical help required?

Only going to happen if you’re playing at Onyx level.

Ranked needs to rank you as quickly and as efficiently as it can… so it can get on with it’s job of match making.

Which group of players do you suggest they use to boost your ego while you wait? How will this help them to become addicted to the game?

I thought it was really good blog actually - very informative although I’d seen similar info on here before from members of the community.

It’s good to see they recognised community feedback and questions with that faq section - proof that they are listening.

2 Likes

I appreciate them coming out and giving us their logic and how it is supposed to work. Genuinely, thanks 343. More of that kind of openness will go down well with the community I feel.

As for the actual content of it…. Sure, most of it is logical. And mathematically it makes sense. But that does not translate to a good experience for the player in my opinion. I personally do not like this idea of a hidden hand pushing you up and holding you down based on your perceived skill. It just makes the game really unrewarding and makes it FEEL unfair.

People are playing this to have fun, and to feel like they are rewarded for their performance. If you play a killer match and go 30-15 in a ranked game at your skill level and you get barely any CSR because the system has decided you are at your peak, then how can you blame the player for feeling unrewarded. There should not be HIDDEN barriers or HIDDEN leg up, as it creates a sense of unfairness to the player, whether mathematically right or not. I don’t disagree with anyone saying this system probably gives a reasonable estimate of skill over the long term, I just note that it is unrewarding and alienating players. Just my two cents.

2 Likes

It’s not so hidden. Your CSR essentially mirrors your MMR.

And the “hand” pushing you up and down is your level of skill.

Nothing unfair about it.

Which is why an XP rank (weighted to performance) odd long overdue.

If you want to rank up you need that kind of performance above your skill level.

One good performance at your skill level doesn’t really tell the system a lot.

The same players who want that game ‘rewarded’ would just complain when their next few games are against harder opponents.

That’s literally not true. You’re lying.

MMR does not mirror CSR because you can tank your MMR in Quick Play and then go into rank and you’ll get easier games. On the other hand you can go into Bot Bootcamp and get matched against Onyx players because the games perceives you as being really good when you just played bots.

I am currently sat at unranked on all ranked variants. Before the reset I was a platinum 5 (I think, maybe a 4 or 6). Since the reset I’ve played 6 games and have lost 4 (again, I think, can’t remember) and have been against Onyx players in every single match.

The system is fundamentally broken and does not work.

I’m pretty good, but I’m not Diamond good let alonme Onyx. The system needs a complete overhaul to make it fairer and easier to play ranked. Currently, ranked is a chore and pointless.

2 Likes

I’m not saying it is a bad ranking system, but it is an unpopular one for a reason. I don’t disagree with most of what you’re saying, I just don’t like the experience it creates.

As for the hidden thing. It is absolutely hidden. You cannot see your MMR, it is obscured. And it is the MMR that drives your rank. That is the definition of hidden. Whether it is mathematically correct or not, it is a hidden hand that guides everything, and that is the reason people do not like this system. It’s a perfect example of why applying a mathematical model to everything does not always go well, because it ignores the human factor.

My CSR after my first few placement games was like Plat 3. I ended up at Onyx before the reset, but played high diamond and onyx players throughout my entire ranked experience, all because of this hidden hand in the background. The CSR was meaningless. I have no motivation to play ranked again until next season.

Lying is a strong opening statement.

It clearly does.

I mean. Seriously. One of the biggest factors in how much CSR you can gain/lose is when your MMR and CSR aren’t matched. The MMR is always pulling your CSR back to its level.

It’s literally how it works.

Nope.

This only works on new accounts and you go into ranked for the first time. In this situation a seeding MMR is assumed from social or botcamp.

Doing it with an established account is much less effective.

If you look at there original Reddit threads there were plenty of posts from people who tried and it didn’t work.

Otherwise we would have countless YouTube’s of Bronze players showing off their Onyx ranks. No? Because it doesn’t work.

Your MMR is taken into placement.

And it’s not unusual to have some tough games early. The system increases its uncertainty in your rank and gives you a chance to rank up.

This is a good thing.

The important thing is that you play accordingly and you come out with an appropriate rank.

It won’t be long and the system will be confident in your rank. Your opponents will be more consistent. But it will also be harder to change your rank.

All games you play impact your expected performance (MMR), which in turn impacts how much you gain or lose in a ranked game (CSR).

In theory this sounds good. In practice, it’s really dumb. Essentially, everything you do (social, BTB, fiesta) impacts how the system rates you as a player. So when you fire up ranked, it uses that as a baseline to determine your elo rate (up or down).

The MASSIVE problem with that is, people don’t play the same with social playlists as they do ranked. For example, I play with friends (sometimes having drinks) when I hit up social and BTB. I play ranked very differently, as do most people. Can someone explain why my performance in a fun, casual mode should have any bearing whatsoever on my rewards in a ranked system?

All that does is make me (and others) have to sweat in social, since the system tries to match the CSR to your MMR. MMR being an accumulation of skill across ALL game modes.

This essentially removes a safe space to mess around and instead makes everything an extension of ranked.

Ranked should be it’s own thing. CSR vs CSR. Nothing else. Simple. If I’m 1500 and beat 1600 players, I should go up, a lot. The inverse should be true too. Adding that superfluous extra layer of MMR to ranked serves no purpose at all. Other than to obscure what’s really going on. Which is obviously what’s leading to the negative feedback.

The solution is simple - CSR vs CSR. Stop trying to outsmart your players.

4 Likes

Their blog post was useless gaslighting propaganda that didn’t address anything.

They had a pre-scheduled CSR reset for what was originally going to be the launch of season 2. They saw a PR opportunity to say “this is fixing things” but basically nothing changed.

None of this system’s fundamental problems were even acknowledged. Their system, which matches everything based on a global hidden MMR that everything in the game affects, is absolute nonsensical trash and they refuse to listen to that kind of feedback.

This is one of the (many) examples of 343 making core decisions about their game that are hilariously and obviously terrible, but they are unwilling to eat a slice of humble pie, admit it, and actually improve it.

2 Likes

The blog wasn’t bad but I’m VERY disappointed and not happy at all to hear the social play affects your MMR. That is RIDICULOUS! Social play is just that, social! I’m going to try and do things in that (like load up in a razorback) that I would never do in Ranked. Now that I know it affects it, I don’t care how little, I guess I need to lay top notch all the time.

A other thing I “loved” (not really) is how they didn’t even say everything that affects trueskill2. They said "TrueSkill2 uses a variety of data about each player’s performance in multiplayer matches (including wins, losses, kills, deaths, etc.)

So what else is it besides wins, losses, kills and deaths? Do assists not matter like in Halo 5? Does damage dealt contribute? Shooting percentage? Instead of saying etc why not actually name everything that affects it.

Like I said, I think the blog was good overall and appreciate the effort they gave, but honestly I’m more frustrated and pissed off with learning and the confirmation about social affecting MMR and not saying what EXACTLY contributes to the trueskill2 system.

This whole thing makes me NOT want to play as much at all unfortunately. This game just has so many issues…Very disappointing. makes me really sad :cry:

2 Likes

I thought it was ok. Confirmed a lot of what many have been saying. One major thing I thought was lacking was clarifying if your performance in social affected your matchmaking in ranked. I know it said that it uses social to help when it is lacking data, but what about when you’ve got a solid base established? Are there 2 MMRs, one for ranked and another for everything else? I think there should be. Clarity on that would be nice.

1 Like