How is it that you lose a game your supposed to and your rank goes down more than it should? You also win a game that you shouldn’t, but gain less. This ranking system is so backwards
1 Like
If every game is matched with the same average MMR or there abouts why are there games that we should or shouldn’t win!
It’s a total conundrum.
I have no problem with ranking systems that award more or less points according to the rank of your opponents, or some degree of personal performance.
After all, that is how systems like TrueSkill work. It’s literally why they are so much faster at working out your rank compared to simple W/L ladder sorting algorithms.
But it is confusing (and frustrating) when you can’t easily work out what just happened. Especially when the moment is still emotionally raw.
And it’s even more confusing when your CSR moves differently to your MMR.
They really should have an option to click on your CSR bar and see a break down of what just happened. A little report card that lists the factors involved that resulted in how much you gained or lost. It doesn’t have to be exact numbers… just an idea of what is going on.
eg. Your CSR went up 8/15; Opposition rating -. Personal performance ++. Your MMR is unchanged and currently below your CSR.
And then people may be less angsty towards the system as a whole.
In the matches I’m referring to the problem is that I play well and it’s a match that the odds are against you, which is why I’m bringing it up.
There were discussions with Menke about those types of things being in Infinite. I wonder how much of his plans got scraped because of developmental constraints occurring everywhere during the game’s foundational work?
You would think the time-saving default would have been to use the same weightings and parameters as H5 (ie. what Menke already had set up).
The system probably lacked confidence in your rank (your MMR curve was still wide)… and/or your CSR had drifted above your MMR (on the back of wins vs lower ranked sides) and it used this match to claw back some CSR.
But the bottom line is we need access to a breakdown of CSR gained or lost.
1 Like
Oh, I’m referring to the extra analysis of the CSR adjustments. He was talking about all kinds of additional data to help people comprehend their rank and its adjustment post match.
2 Likes
Ranking needs to be a mix of personal performance and win/loss. Not solely win/loss.
1 Like
As always - the problem is how to gauge performance.
Even in Slayer there is contention over raw kills vs kill/death vs how to incorporate assists. And that’s before we start up discussion on the potential for damage stats.
Then Objective games. What a minefield!
And even if you do decide on a particular metric - eg. planting the flag, how do you stop it from becoming toxic? I’ve already had players shoot me in the back metres from home, just to steal the flag cap.
I am more than happy to make it all about the win / loss. It may not be as satisfying on a game by game basis - but overall, all the little things you do that contribute to the win (or loss) will come out in the wash.
Easiest way to explain.
Visible Rank - 1,000
Hidden Rank - 800
These can differ for many reasons such as you win a game but your performance was below expectations so your hidden rank drops, even though your visible rank increases. It is also affected by other playlists and time decay when you aren’t playing.
You are matched on your hidden rank, so a player who is truly a 1,000 rating will dominate insanely Vs 800 players and win most games no contest. If you can’t then the system realises you should be lower. You are being matched with lower rated players so if you are really worthy of a better rank then the games are significantly tipped in your favour.
Your visible rank moves towards your hidden rank. So even with a loss your 1,000 might drop to 950, but your hidden rank increases to 850. The ranks are getting closer together and even on losses you are making progress towards a higher hidden rank even if your visible rank is no longer appropriate. At a certain point these numbers even out and you will either climb, stay or fall in rank according to your regular playing ability.
Short version, if you are losing lots of rank even in unfair games then your hidden rank is much lower than your current visible rank. That won’t keep happening the more you play as it evens out over time. Your peak is rarely your true rank.
Ranked system in this game is terrible. Just go negative in bot bootcamp for a few hours if u wanna rank up.
That’s a fallacy.
If you dominate on bots with a brand new account you can seed your ranked MMR and at least start off with an insane rank. But you will be miserable placing against high Onyx players. But again, this only happens with brand new accounts that have never entered the ranked playlist.
If you already have an established rank, then doing badly in bots may get you a softer game or two (there are shared weightings for form etc) - but because your ranked MMR is resistant to change it won’t allow you to rank up. Any small CSR gain will be lost quickly.
There are easy metrics to use in order to stop people from stealing flag captures. Example people within a certain radius of the flag carrier would be considered “defenders” and could get an assist equal to the same points of the flag capture.
At the end of the day adding more personal performances metrics into the game just incentivizes playing the objectives which will in turn help teams win. But it should be an even 50/50 when determining rank granted or loss at the end of each match.
It would also help games not be extremely sweaty.
At least once my “team-mate” stood by and clearly waited until the enemy had killed me so they could swoop in, get a kill and a flag cap.
So I don’t underestimate how far people will go to milk a system.
I agree that incentivising playing the object is a big win. I just worry how it may end up promoting toxic play.
And if you start complicating by having multiple rules about who was standing where and when, holding what, and looking in what direction? It’s just so many layers.
And yes, sorry. I am exaggerating a bit 
What about the people who move in behind the flag carrier and defend. It’s a sacrificial act that sets up the flag cap - but all they get is a +1 in the D column.
The person who was first to the flag to return it. They often die just before someone else swoops in and completes the return (often with an easy kill to clean up).
The sacrificial act of planting a stick grenade on the enemy with sword?
The person who grabs the oddball despite no/low shields and jumps it out of bounds?
A win is made up of so many acts - how do you possibly judge them all.
I guess that is why I am happy on the win and the loss. It’s a team game. And if you keep doing the team things - you will win more than you lose and rank up appropriately. Everything you do all comes out in the wash.
And really, once the system has found your rank - the games where you shine as an individual don’t tend to happen. Pretty much everyone on your team ends up putting in the same effort. And if you do have a personal day out, it’s usually because you are playing a team ranked below you - and you won’t be ranking up much anyway.
I know it’s frustrating when you were best on ground but your team lost. But there will be just as many times when you win despite you having an off game. Take the good with the bad. It will all balance out in the end.
1 Like
Well, except not really. Plenty of people, myself included, have inexplicable CSR gain/loss well past the point where we should be stable and have MMR/CSR close.
Last week in consecutive games I lost 11 CSR after losing a 46% win-prob game with a perfectly par individual performance, then regained +1 CSR for a win the next game despite having only 40% win prob, and a strong individual performance. And then another +1 the next game after another strong win.
The theory’s all well and good but there’s just no way to explain what’s going on from a game-to-game basis. As Darwi has advocated, we badly need a breakdown after each game when the behavior is this erratic. Even if it makes sense on a population (e.g., 100 game) level, that doesn’t mean it’s fun or rewarding for players to be jumbled around in the meantime, especially long after it should’ve stabilized. It’s why I’ve stopped playing rank (and Halo) except on occasion when friends drag me in.
Very explainable though. Your CSR is higher than your MMR. Your MMR can decrease outside of the playlist as it decreases over time not played and for poor performance in high correlation playlists.
One game is irrelevant because it’s over all games. Your MMR will have increased by a good performance but will still be far below your CSR, which is why you make negligible gains. To get greater gains, you need to play more and continue to outperform your MMR to keep pushing it higher. As it gets closer to your CSR you will see more even gains and losses, then as your MMR surpasses your CSR you will get huge gains for wins and -1 for losses.
Why has your MMR dropped? You tell me. Consistently poorer than expected performances or time not playing are the main things that come to mind.
Obviously it’s easier to do a good performance when your MMR has already dropped as you’re being matched Vs worse opponents. It doesn’t mean that current CSR is warranted. You have to earn that back. Like I said before your CSR could be 1800, your MMR 1600. For a win your CSR would only go to 1801 but your MMR could go to 1650. Then if you lose with a really good performance your CSR might go to 1786, your performance is good but it’s trying to match up with the MMR figure which is now 1675. You are making MMR progress but not CSR progress which makes sense as you aren’t playing Vs 1800 rated players. You just were 1800 rated once upon a time. Prove it once again.
Keep playing, your rank averages out.
Why is MMR drifting for players like me with hundreds and hundreds of games consistently played? If the game still can’t match my CSR and MMR after all this time, it’s a pretty sad system.
Anyway, my point above remains. People simply don’t like it because they can’t figure out why it does what it does on a game by game basis, even if they understand the theory. Thus the constant threads. Even understanding the system, I can’t make sense of why it does what it does on a game-to-game or even 25-game to 25-game basis, as was the case when your only explanation for my 25 game streak way back with 80% wins for net negative CSR hundreds of games into a season was “guess it’s CSR drift”
But hey. We’re dug into our positions. If you’re still having fun playing the game, power to you.
1 Like
MMR drifts for everyone, because your skill level is not constant. It constantly moves with you and the rest of the world. Your MMR can go up on losses and down on wins.
You were 1,216 and you lost 9 games and won 3 and you ended up 1,213.
25% win rate and a net loss of -3 CSR. Sounds like your MMR is higher than your CSR.
Is the issue that you are far below your peak?
The system exists purely to get players to be better matched skill wise. Technically they could scrap CSR and use MMR only but that completely removes the team play aspect as players could 1) ignore losses if it meant visible MMR gains and 2) could easily work out how to manipulate the system most effectively. Having the obscurity makes it less likely to be abused in any gaps which it no doubt will have.
1 Like
I made the dire mistake of giving Ranked a try again, and promptly lost 15+ CSR for losing a game we had a 6% chance of winning.
Aaaaand I’m out again. That was easy.