It would only equal the playing field if you could prove all problems with the game were happening to everyone, equally, everywhere all the time. Some players play more than others. Others have better proximity to servers. Other players cheat or geo-filter to tilt a match to their advantage. There are platinum who play like onyx, and onyx who play like platinum, due to imbalanced matchups and mass account creation. Some people play at peak hours, others play whenever their time can allow it. Telling people they aren’t as good as they think they are and acting as if CSR is a completely accurate representation of a player’s skill without considering these factors makes your original statement is an oversimplification. There are just too many factors right now to make a statement like that.
This ranking system really is absolutely horrible. I don’t understand how they still haven’t found a way to improve upon Halo 3s ranking system.
They did!
-
Get rid of the 20 hidden ranks.
-
Separate ranking from the XP grind.
-
Put measures in place to counteract rank locking.
But where Halo 3 did get it right was keep the MMR number hidden. People had to be, and were, satisfied with being a ‘50’. There wasn’t toxic play / manipulation trying to get to 50.1
Yea, fair one. It was an oversimplified statement that’s not true for everyone.
Also would like to note that I have read the paper (thanks @Darwi for the suggestion) and overall I have come to the conclusion that even if TrueSkill2 is working as intended, it doesn’t feel good to players. But I also have a hard time believing that it is where it needs to be.
At launch, I was seeing consistent gains in CSR all the way up to Onyx 1908 (yes, yes, I know it weights your earlier games more). But then they did a balance patch shortly before Christmas and since then, it has felt like I am just trying to maintain my rank but it is made impossible to do so because I will lose inevitably… No one can have a 100% win rate.
I think the frustrating thing to me, is that when I am expected to win and do I only get like 1 or 2 CSR gain… But if I am expected to lose and I do, then I lose like 10-15 CSR. As a player, I would expect that if I am expected to win and I lose, I should be punished. But if I am expected to lose and I do… great! The system accurately guessed that I would lose. (And I get it, there is the whole drift up and correct down notion but it just feels bad).
I don’t think this is happening as much as we might think it does though I would love to see stats on it… Lucid (the pro player) has been at ~2500 since like early December or something. And there is no way he is grouping with platinum players or whatever to push his CSR higher. Sure, maybe he is an outlier because he is literally a professional player, but I don’t think that the argument that people are exploiting the system is totally valid. If anything, to me it proves that the parameters 343 have tweaked do not function at the top tier of play (and in my opinion that is basically all Onyx).
Anyway, sucks that 343 doesn’t care about its community and just stays radio silent on pretty much every issue the game has.
Welcome to the club. It’s small… but we’re working on it. 
I tend to agree. The MMR is fine. The CSR not so much.
They did mention the rank distribution was top heavy at the start - so they had to reign it in a bit.
Yep, frustrating as heck. But if you are in a ranking groove where matchmaking is giving you 5-6 wins for every loss (because that is the best it can do with the teams available on the server)… then that’s pretty much how the CSR has to behave.
He would be an outlier because he just doesn’t lose. Freak!
By it’s nature ELO systems are open ended.
Yep. It no longer functions as an ELO system because the good teams are just not playing each other regularly enough.
There are too many games where one team comfortably outranks their opponents.
So they have to artificially limit the CSR gains for a win and boost them for a loss.
Normally you would continue your (albeit slow) ranking journey at a 50% W/L. You would pretty much gain the same for a win as you would for a loss. It would be up to you to cause an upset and claw your way up the food chain.
But if matchmaking can only serve you up 1 opponent in 6 that is better than you - then it has to take away five times the amount of CSR for a stock standard loss.
And the other problem is that this isn’t the same experience for all Onxy players. The ones who are getting more matches vs better teams will fall harder and struggle to move up. While the ones who are getting even more soft games will “break free” of the chains and tend to keep accumulating CSR (albeit very slowly).
This is where I think those teams who are manipulating the match-making (eg. carrying a low ranked Smurf) can take advantage of the system. Grinding out more and more wins vs lower ranked teams.
So what’s the solution? I really don’t know.
The ranking needs better data. Ideally tournament type structures - but that is impractical.
Do you limit games that actually count against your rank to those games where everyone is within a certain range of rank and everyone has a minimum ping? It may create a lot of “dead rubbers” - but at least the ranking would be consistent.
Do you hide everyone’s MMR number and then regularly (eg. weekly) normalise the range back within the normal curve of 0-1800. You would have to make the wait worth it each week (leaderboards, prizes, lots of fanfare, etc).
Do you abandon the CSR number altogether and just leave it at a division and tiers - and instead give everyone an XP rank to grind.
The thread below is getting a lot of attention over on r/halo. Would be really helpful if we could get an official answer on this, but I’m curious what the thoughts are on it here.
https://www.reddit.com/r/halo/comments/sjcg8u/its_official_matchmaking_is_broken_social_games/
I’m not sure that thread is really telling us anything new.
-
We know that the playlists share MMR data (with offsets for each playlist).
-
We know that if you enter a new playlist then the system assumes a starting MMR from a similar playlist.
So, yes. You can bump your MMR up in social or bots and go into ranked with all guns blazing.
But TrueSkill2 will get to work pretty quickly on finding your proper ranking.
It’s telling us the system is hot garbage, but yes, we all knew that.
What interests me moreso is the rapid response on reddit, and how people react after realizing what they’re doing in other playlists is affecting their ranked matches. I personally don’t like it, and it does seem to resonate with many it could be a bad thing. I think we just need 343 to come out and explain to what degree this is affecting things.
I’m also curious on the validity of whether or not the system is asking people to carry based on MMR or be carried based on bad past performance. I saw this over and over again in the thread, and heard it recently in a LucidTW clip. I will say it often feels like the system is doing this based on my own experience.
I agree that more info from 343 would be great.
But outside of the influence of a seeding MMR for a new playlist (eg. the Bot scenario) - I haven’t seen a lot of evidence.
TrueSkill2 does try and take into account form. Originally it didn’t (your rank was expected to head in either direction equally). TrueSkill2 makes the assumption that your rank is more likely to increase (the walk is positive) - which helps it to rank you faster. There is also a weighting if you haven’t played for a while - which I guess also works as an effective positive influence if you are playing more regularly.
But again… some sort of FAQ / feedback from 343 would be awesome.
But again… I think all these things are relatively minor distractions. The main problem with the ranking is matchmaking vs the CSR. If the system can’t give you 50:50 games then the ELO system breaks down. Your winning streak vs lower ranked teams will allow you to rank up and up and up…
So if you get 1:5 games vs teams above you they have to cut the CSR for a win and increase it for a loss to keep CSR = MMR. This gives you the frustrating CSR oscillations around your actual rank that we are seeing.
It also makes it tough for teams that are good, but not good enough to win all those expected games… and makes it easy for the very good teams who go on to grind their MMR way past the normal distribution “ceiling” of 1800.
Looks like Rocket Sloth just made a video on it: “Why your rank in Halo Infinite means nothing.” Not a good conclusion to come to.
https://www.youtube.com/watch?v=y44ApR-O30Y
Yeah, ranked at this point is nothing but cheaters faking their MMR with a low ranked teammate.
Edit: I stopped counting at 10 games in a row with that type of player in it. Not even sure what I’m at now.
1-50 Ranking system from Halo 2 or Military ranking from Halo 3 is all this game needs
There should be a reset at the start of every season. This allows for better representation of player skill, and allows for updates to be incorporated to game modes and playlists.
I don’t believe getting a high rank for a great player should ever be a grind though. 10 placement matches is enough to say, “okay this guy obviously isn’t a gold player, so he shouldn’t be matched as one.”
The logic needs to be tweaked when season 2 drops. As others have said the current inner workings of the ranking system seem to more heavily weigh individual score than wins, and this really shouldn’t be the case.
The difficulty I have with such posts and videos is that they try and polarise the conversation.
Ranking has some issues. For sure.
It doesn’t mean it is completely broken. Or even broken for that matter.
It doesn’t mean that, for the vast majority of people, that the rank they currently have isn’t an accurate reflection of their skill. Even if their CSR oscillates around it and frustrates them to all heck. Like it or not… their rank is likely real.
Yes, there are some people out there manipulating their CSR and grinding it higher than they should. But every system has those people.
343 just need a chance to reflect and revise. Baby and the bath water.
Kind of.
They certainly need to change the CSR to simple ranks. Showing the actual rank number is counter productive and increasingly inaccurate.
But 1 to 50, or Bronze to Onyx, of Private to Bearded guy with lots of medals… it doesn’t really matter how you represent it.
We’re assuming that every season will have placement again.
You don’t need a hard reset. You can take your previous MMR into placement. You just need to increase the sigma value and make the ranking volatile again. This will give any player who is underperforming a genuine chance to rank up quickly.
If you reset the MMR you are just going to make it Smurf city.
But having said that they look to be updating and resetting the system very soon.
The whole idea of ELO systems such as TrueSkill2 is that finding your rank isn’t a grind. In 10 games they can ball park you. In 50 they can nail you.
The grind is improving your rank.
Once you have you have hit your ceiling it is very hard to improve. This is true with any skill, from Halo to crochet.
I think the evidence for this is more anecdotal. Nobody ever posts when they played badly but were carried to an upset win and earned a big chunk of rank. The emotional focus is when you perceive someone else getting more than their share - even though you have no context to their CSR journey.
As of lately I’ve been noticing while solo queuing that at the start of a match when a teamate leaves / crashes I’m still being penalized for the loss. It’s not a big loss, but its usually like 3 CSR.
Problems:
- People quitting
- Unranked people in matches
- Throwing a Platinum or Diamond into the mix
How is this fair? 3rd time in a row someone quits right at the start of the game. What’s the point of playing if the odds are most often stacked against me with both people quitting early and having multiple unranked people on my team? If I play the objective and have a low KDA, I’ll rank down more, right? How is that even fun?
The biggest problem are people quitting, or people dropping especially before the 1st match even starts. And it seems like I’m ALWAYS playing with unbanked people. Is it even really a ranked arena if every games has unranked players on either team? It also seems when I play only ranked people, which is been getting rare, that I don’t get caught in a situation where I feel like it’s only up to me.
I was Gold 6 before the rank reset and I got placed at Gold 1 after. What is even the point for me to play ranked if I’m going to keep getting ranked down because of more games than not of unbanked players and people quitting early, only to have to start all over in a few weeks and problem have to work my way up again.
It’s sucking the enjoyment and feeling truly challenged with some really great games when all people know what they are doing and are a similar rank to me. Not having to get on the headset and explain how Strongholds works over and over and over.
Ranked Capture the Flag
My team:
- Gold 5 (me)
- Unranked
- Quit early
- Gold 1
Other team:
- Platinum 5
- Silver 6
- Gold 3
- Silver 4
Ranked Oddball
My team:
- Gold 5 (me)
- Unranked
- Quit early
- Quit early
Other team:
- Silver 3
- Gold 2
- Platinum 5
- Gold 2
what i have a problem with is leavers they leave and stuff like strongholds becomes very hard and there is no way to avoid that period like if i want to leave cause its over since they can two man rush everytime or three man i cant just leave cause of csr loss or i lose and get csr loss there needs to be a fix for that maybe just no csr loss for the team that loses with a leaver idk just thinking