What is wrong with this ranked system, it is Absolutely insane. I win multiple matches in a row and barely see any progress, yet one loss takes a massive chunk away.
I can’t believe this garbage it is infuriating.
5 Likes
Completely mystery my friend. Just about everyone who plays ranked has this problem.
2 Likes
It’s how ranking systems work.
If you want to rank up you have to beat a team ranked above you. If you don’t want to rank down then don’t lose to teams ranked below you.
When you reach your skill ceiling you’ll fall into an oscillating pattern of inching forward (the game tries to give you a little bit of CSR for the win) and then falling back with your next loss (your CSR is bought back in line with your MMR).
I agree it’s frustrating and that they need a better way of presenting your CSR journey.
But it’s likely that your true rank is around where you’re at.
2 Likes
Ranking systems work best in tournament type structures. Each game provides rich ranking data to the system.
The problem with computer games and match-making is that you can play 20-30 games at the drop of a hat - and most of them aren’t going to provide a lot of valuable ranking data. Especially if you are towards the top of the food chain on your local server.
Can 343 just sticky a post or something. This comes up so often.
2 Likes
Ideally they would have a FAQ of some sort. Or a blog post.
Something.
I did think about starting a reference thread about the ranking system. Collating what we know etc. And that way I could just link back to the thread as an explanation.
But I fear it would just be a dumping ground for toxic posts or people calling me a 343 shill.
2 Likes
Or noone will read it 
1 Like
A reference thread without any knowledge to base it on seems like a poor idea unless one wanted to focus on really broad strokes.
The only knowledge we have about TS2 is based on a four years old paper that was developed for a different game. We do not know whether parameters have been changed, added, etc unless there’s information you have that others don’t.
If you’ve played enough ranked games there’s clearly a lot of parallels between the paper and how Infinite works.
Just looking at players profiles validates a lot of what we already know.
There are of course many undefined things, but there is lots that can be said with certainty.
And I guess we’re mostly talking about covering the basics here.
Repeatedly.
eg. “I’ve just won five games in a row and I’m not Onyx”.
eg. “Halo 3 was perfect”
eg. “Bot games are ruining ranked”
343 have done a very bad job at explaining / selling the ranking system.
1 Like
And yet, my entire fireteam games has pretty much stopped playing because we realized we just can’t connect our individual performance or what a reasonable person would expect the match outcome to be (based on the other team’s CSR compared to ours) with what our CSR does on any given game or even across a session.
It’s anecdotal for sure, although based on the number of threads expressing similar frustration and the rapidly dropping play numbers, seems it’s not exactly an uncommon occurrence even among people that understand the theory behind TS2. Without a peep from 343 as Darwi noted, the only thing any of have is our anecdotes. Maybe you’re having a much better experience, and I’m glad if that’s the case.
@Darwi, well, that’s certainly fair enough.
1 Like
I’d say you are barely going up because you aren’t killing enough. Game doesn’t actually care about w/l, cares about kills per minute. The longer the game, the more kills you are expected to get. If it cared about w/l I’d probably be like gold lol.
1 Like
Actually, I did take the time to look at my last 25 matches, and would be genuinely curious to get your input @Darwi.
My last 25 matches are a 52% win rate. During that span I lost more than 100CSR to drop from Onyx 1582 to D6 1478. My KD was .98 so my individual performance might have varied but we can pretty confidently say I wasn’t tanking. Onyx 1582 was a little high for me but I’d been Onyx 1530-1560 for months. What do you think happened on this run?
The only game that jumps out is a -15 CSR loss we had against a team with a 86% chance of winning thanks to a silver smurf on their team. There were at least two other smurfs across those 25 games but since we didn’t lose, they didn’t cost us (although obviously we didn’t gain either, it’s a downside only situation).
50% - Loss - 23/27
50% - Loss - 11/20
50% - Win - 5/3
50% - Loss - 9/10
50% - Loss - 15/22
50% - Win - 19/15
50% - Win - 19/12
50% - Win - 14/12
65% - Loss - 16/20
88% - Win - 14/11
40% - Loss - 9/12
43% - Loss - 11/22
56% - Win - 13/8
70% - Win - 21/26
56% - Win - 17/10
86% - Loss - 20/24 (Derankers are total losers)
50% - Win - 17/20
50% - Win - 22/21
50% - Win - 19/4
50% - Win - 19/16
50% - Loss - 9/17
50% - Win - 7/8
50% - Loss - 12/10
50% - Loss - 21/18
50% - Loss - 12/15
I’m using k/d only. I’m not gonna analyse every player, score, damage etc etc. There’s more that goes into it but here’s a rough overview from my understanding.
So looking at the games the 25 started with nine 50% games. These games can go either way but something that also needs to be considered is that in these games a players performance is not expected to be equal to their teammates, it’s supposed to be comparative to your rank and everyone else’s rank. So roughly looking at your friends I think you’re second highest rank? So technically you’re expected to do better than the other two as a higher rank. That’s how two players can play every game together and get different CSR gains and losses. You may be expected to get a 1.2 k/d in slayer because they are expected to get a 0.8 k/d. If you get a 1.1 k/d you are underperforming, but if they get a 0.9 k/d they’re overperforming against their rank.
After five games, your MMR could already be lower (it also could be lower than your CSR before these games), this means your opponents are actively getting worse. You might be winning games but still not performing at a level that would increase your MMR, beyond where your CSR currently is at this point. You won the next 4 with 3 games of mediocre k/d and one extremely good one 19-4, probably boosted your MMR back up a little. Bare in mind your MMR may still be lower than your CSR. This means you’re fighting an uphill battle. Your opponents are steadily getting worse but your k/d and win ratio is not improving against worse opponents which means the system is identifying that your MMR shouldn’t increase as you aren’t beating these worse opponents to a level they can identify a higher associative MMR.
That deranked game would have likely hurt your MMR a bit too. What a bunch of losers.
Your MMR can decrease even when you’re winning games which makes the wins give less CSR and the losses even more impactful. It is theoretically possible for a player to win 20% of their games and end up with a higher CSR than when they started based on their MMR. Your MMR can increase off of a loss too. Conversely a player could win 80% of games and end up with a lower CSR at the end of it.
Realistically as your opponents steadily decrease in ability (as your MMR drops so does the people you match) your performance should naturally become better because the games should be easier. Based on your wins and k/d over the next games (particularly the 50/50 ones that hold the most weight) it would suggest the system was correct in lowering your MMR as your new opponents are still proving to be troublesome for you.
In terms of how to increase your MMR most efficiently, we don’t know although k/d is obviously a large factor. The less we know the better as players would likely abuse the system if all was revealed. Just try to have fun with it. If I played a night with my friends and won 52% of our games I’d consider that a good night of gaming, regardless of where the rank is at the end of it. I hope you and your friends can look past the rank and enjoy these close and competitive games. Over time you might find your rank naturally drifts upwards.
2 Likes
343 definitely need to step up their game. The CSR “journey” is frustrating a lot of people.
Personally I don’t think it can ever scratch the itch that people want it to. It’s not a number you can just read as a value (interestingly the system never purports to know your actual skill level - it’s just a curve that your skill level lies somewhere along).
There are too many factors involved spread over more than just the game you played.
I wish they would drop the number and just leave you with a divsion / tier.
It is all about the wins. Or specifically it’s all about the wins that matter. You need to nail the ones vs better teams when they are up for grabs.
The kill rate is a weighting, sure. But that’s all. It identifies the big fish in the small pond and speeds up their ranking process.
As you would expect, as you match games vs equal opponents, the kill rates drop across the board.
I think part of the fallacy is that people see players ranking up faster than them and assume that their higher kill rates are the difference. It’s more that they are better skilled and have a higher trajectory (because they are beating better teams). Their higher kill rate in that particular match is just a reflection of their skill.
It looks like @Ken_CTF has already had a look.
But kudos for having some insight into your own rank - when you said 1582 felt high to you. It’s rare to see someone who isn’t jumping up and down and accusing the game of withholding higher ranks from them.
Maybe you scored a smurf or two as team mates!
1 Like
Last night I won 4 games in a row at Onyx level for a total of +7 csr. I then lost a single game for -14 csr. What a complete waste of time to be lower than when I started after 5 games because of a single loss.
1 Like
A big part of the problem is that 343 unnecessarily confuses the system for players by the way they visualize the progress.
A friend and I were playing the other night and we were checking the progress of players after one of the games. On his screen it showed the person losing rank while on my screen it showed the person gaining rank.
I’m not sure we can even trust what we are seeing at this point.
From a ranking point of view you are still at or around your actual rank.
Winning 4 games in a row doesn’t guarantee you any increase in rank. Those wins were obviously expected - hence the token CSR gain.
The loss was either unexpected or your CSR had drifted higher than your MMR (courtesy of said token wins).
But essentially you are still at / around your rank. I mean plus or minus 7 CSR is hardly a significant change in either direction. Have you done anything to justify ranking up? Has your play/skill level improved?
More importantly I hope you had some fun in those five games. That is the important determinate of not wasting your time.
I think they recognised that the natural fluctuations in CSR would be frustrating - so they tried to hide within the “CSR change per session”.
Weird. Maybe just a server delay - one of you got the old info and one the updated one?
I understand what you’re saying from a technical perspective, but the player experience must be taken into account. If players can’t reasonably feel like they’re making progress when they win, they won’t continue playing because it’s a waste of their time.
1 Like
Super helpful and involved breakdown. Thank you for taking the time. A few thoughts and reactions, and I acknowledge my obviously defensive bias–
-
The differences in rank with me and my squad are not that severe. It’s high diamonds and low Onyx, so I don’t think there’s not a crazy expected difference in performance between us. Maybe at 1582, but not at 1550 or 1530, for example. We’re talking small margins so your comments about how I’d be expected to perform substantially better feel questionable. Me vs. a D6 probably shouldn’t have a .4 expected difference in K/D, to use your .8 vs. 1.2 example. But maybe I’m wrong.
-
You describe how I may have been losing MMR during wins. Can you elaborate? My worst K/D (I know it’s a bad metric but we’re gonna stick with it) was .81. Yes, it’s not incredible, but is the suggestion that a team win in a 50% objective game with .81 KD can still result in a loss of MMR? At that point, the trolls are actually right, and we should completely ignore W/L and just slay, because that just reads to me like TS2 ignores win/loss which I don’t -think- is true.
-
Also, feels like you and @Darwi are at odds, as he noted in the next post that “it’s all about wins [qualifiers about expected win pct, etc]”. In fact, looking at the odds, if we throw out the smurf, we actually only lost one game out of 25 we were expected to win. And we won more 50% games than we lost. So, on a purely W/L perspective factoring in expected outcomes, we did perfectly well, but our CSR was in freefall.
-
Yup, we tilted at the end of the last session. Every time we won, we gained 1 CSR and every time we lost we dropped 15, and it led to some poor play. No way around it. However, we didn’t actually get weaker opponents later in this 25 game streak compared to earlier (as far as CSR), which punches a hole in the narrative that we couldn’t keep up with the opponents as they got worse (thus validating the MMR drop). Our last 3 games look like they still featured roughly the same mix of Onyx and Diamond (except perhaps the very last one), and some matchups even look tougher than previous matches considering we didn’t have our other Onyx player so we were a relatively weaker squad.
-
I appreciate your sentiment about ignoring the rank, but we’d play social if we didn’t care about rank. It’s the whole point of the playlist… it’s in the name haha.
Side note, why do some games have expected win % on HaloTracker, and others don’t? Also curious how that’s calculated, some sort of homegrown thing based on CSR, I suppose?
1 Like