I guess I’ve been lucky. Haven’t really noticed anything different in the last week or so. The only matches that seem to blow out are those vs CSR grinding stacks.
Looking back over your Halo Data Hive history;
41 v 50 (after you quit) = close enough
199-200 (after you quit) = very close
50-48 = very close
50-36 = one sided (opponent quit early)
50-40 = close enough for social
40-50 = close enough for social
27-50 = one sided (again, after a quit)
and then a series of land grab. I only looked at the last seven on that page of results; 11-9, 11-8, 11-8, 11-7, 11-7, 11-6, and 11-3. I haven’t played a lot of Land Grab but I would have thought only the last two were badly matched?
So there were two very uneven Slayer matches - both which could be blamed on someone quitting. And one very uneven Land Grab (3-11).
Am I missing something? I wouldn’t think those were a spread of scores worth shutting up shop over.
Ranked Matches tend to be more unfair than you’d hope for regardless if MMRs are close or not. If anything the MMR seems to be less important in the amount of blowouts I see in Ranked.
I do experience more close matches that I did in Season One… BUT I still experience many blowouts either in favor of or getting destroyed on.
I’ve experienced many matches where MMR is close but the ranks are one team having Onyx Players and High Diamond players while the other team has maybe 1-2 high diamond players and 2 platinums or in some terrible cases a gold player who has no business playing high diamond or Onyx players.
I’ve also experienced some matches where I face a team that is a squad of 3 (or 4…) and they are all high onyx players, and I might get paired up with maybe one Onyx player (if lucky…) and maybe a Diamond 2 or some Platinum player. You’d expect the system to have a squad of four onyx players face another squad, or at least try to pair up all high diamond / onyx players for the opposing team.
I don’t 100% know if the system focuses more on MMR or on Ranks for Ranked matches…, but I don’t think there should be any reason a Gold player to be in the same match as an Onyx player 9 times out of 10.
And it can sometimes be hard to read. You may have someone with a CSR much lower than their MMR, like just after placement.
As for the Onyx-Gold match ups. That’s friends playing together. But increasingly they seem to be organised to try and push the opposition ranks down. I believe it still creates 50:50 matches for them… but it helps those Onyx players grind CSR, albeit very slowly.
While I have seen some golds play with friends, I’ve also seen at least once or twice a gold player solo qued get paired up. (Yes I actually look and see if they are a squad or not on my team or the opponent if I get crushed).
I also don’t really understand the point of CSR or Rankings in Ranked if MMR is the placement system anyways.
The match making system will start to push the envelope when players have been waiting too long.
And it’s not silly. It can “look ahead” and see if players are going to be available in the next few minutes. So it makes the hard decision.
The key here is to look at the average MMRs. If they are out of whack the system has to do something.
Consistent movement. The MMR can jump all over the place early. So the CSR can be made to move slowly (max 15 pts per game).
Consistent reward. CSR if made to always go up with a win and down with a loss. MMR can (sometimes) do the opposite.
The other benefits of separating them aren’t really used by Halo. Although they should be. Like having a different scale (eg. 1-50), or protecting against oscillations (delaying dropping back a level until you’ve lost a few games).
The points are odd. Was on an alt that’s an onyx. Lost a game, average team mmr was the same & it dropped me 10 points even though going positive. Since it’s an alt, I don’t mind too much but it can be frustrating at times lol. Dropped 36 onyx points in 3 straight losses going negative in just 1 match, but it was like a -2. The grind continues
Don’t think of it as a grind. More of a ranking journey. Just play the best you can and your rank will take care of itself.
Trying to micro-analyse the CSR change on a per game basis will just frustrate you.
The usual cause for dropping a chunk of unexpected MMR is that your CSR had drifted above your MMR. You get a few wins vs teams at or below your level and CSR goes up but not MMR. With your next loss the system uses the opportunity to pinch it back.
343 could fix this by reducing the scale of the CSR. Not saying they should go back to 1-50… but maybe something like 1-100 or 1-500 would be less jarring.
CSR doesn’t look at your performance in the game.
Really does look like your CSR has drifted too far above your MMR.
It’s pulling your CSR down hard on each loss.
And again, CSR doesn’t look at your personal performance. That’s MMR (kpm and dpm).
In my experience with playing ranked on ANY first-person-shooter titles is that MMR will decide the ranks of ONLY your teammates and then the enemy team is usually above you.
For example, in Siege I am a Gold 1. I rarely enter Platinum rankings.
Mostly because whenever I get into a match I get teammates that are in the Bronze-to-Gold range with maybe a slight chance of a Platinum player on my team.
Meanwhile on the enemy team I can guarantee that two-to-three of them are Platinum or Diamond regulars.
And back in OPERATION // NORTH STAR I saw Champion Weapons charms and icons on my enemy players in a total of fourteen separate matches.
And what makes it all the more worse is the fact that MMR is supposed to restrict you from playing with certain people. 1,000 MMR window for players to be paired with you. If they are outside that range, you should’nt be paired with them.
And yet
Somehow
I constantly get paired with people well outside that range.
MMR is a joke that only exists for bragging rights in video games and nothing more.
MMR in halo infinite is a flawed system. MMR should only be used for smurf detection and once ranks are assigned, players should be matched by rank as close as possible while balancing the matching time.
Big news! New Team Doubles ranked playlist…
Now is this playlist going to use 1 singular MMR for matchmaking? If so, what if people perform way worse in doubles than Arena? It’ll incorrectly match. Same can be said if they add snipers or swat as a ranked playlist. You don’t play the same in every playlist. Not good
It isn’t 343, per se. It is the Trueskill 2 algorithm.
If you already know this, sorry for posting it again. For those that don’t, but should if you’ve been on this forum, or any forum talking about Halo longer than 5 minutes, this should be helpful…
How does TrueSkill 2 work?
“The algorithm will crunch the stats of everyone in the game and come up with a probability of one team winning. So if the game says there’s a 60% chance of your team winning, and you win, the amount of MMR you gain will be cut because you were “supposed” to win that game.”
So, it doesn’t force a loss. It will, however, pair you with less skilled players if the stats it records from your own history seem like you should be playing against better skilled players. The players you get on your team will/may be less skilled than you. The players on the opposing team will/may be in a league above you.
Thus, you lose your next game, or games, depending on how quickly the “crunch of your stats” gets back to where “TrueSkill 2” thinks you should be. Most people will win 2 or 3, then lose 1. Repeat. This is why players have construed this to being pigeon holed in a single rank; Example: Plat 3, and never getting out of it no matter how good they do, or how many games they win.
343 didn’t create this algorithm. It was flawed from the start, but 343 decided to use it because it was probably easier to implement in-game then creating their own ranking system. Sadly, it just makes MM a sweat fest. It means I have to play at, or near, my best every match or I won’t be enjoying my experience. Unfortunately this isn’t feasible for any human. Which is why most people don’t, and never have, liked “TrueSkill 2”.
Technical point. But it’s not TrueSkill2. That’s just the ranking algorithm. It’s not part of the match making algorithm.
Pretty much.
That’s how ELO type ranking systems work.
It’s why they are so powerful.
But the key is that the % to win is not manipulated. The MM system takes the best eight players available and sorts them into the two most equal teams.
Correct.
Um. No.
It takes the best eight players available, and sorts them into two even teams.
It doesn’t look at your “history”.
If the population is solid there won’t be opponents “leagues above you”. But one of problems recently is that the population isn’t all that solid. You still get even teams - but the skill gaps between team mates can be frustrating.
No.
TS2 doesn’t “think” you should be anywhere.
That is in YOUR control.
And if anything, TS2 actually assumes your skill walk is “up”. So if you aren’t ranking up then you really have reached your skill plateau.
Not true. That would give most people a W/L ratio of up to 75%. And that’s not the issue here - people are complaining of being 50%.
It’s true that you tend to grind your CSR up over 2 to 3 wins and lose it all in 1 game. But that’s different. It’s an issue with the way CSR is allocated.
Said player is stuck in Plat 3 because that is their skill level.
It does matter how “good they do”. If they start beating the Plat 4 sides and stop losing to the Plat 2 sides - they can, and will, rank up.
TS2 was created by Microsoft.
343 use it because of their affiliation with Microsoft… and because it is pretty much the fastest and most accurate ranking system out there.
If we are talking ranked… isn’t is supposed to be?
If you mean Social? That’s on you. Lots of people can relax and play differently in Social.
Which is weird.
Ranking systems are fascinating. Particularly ELO based ones.
And all TS2 purports to do is rank you faster than simple rank sorting methods.
I would argue that “most” people don’t, and have never, really read anything about TrueSkill2.
People may not like how the CSR is displayed. Or the current implementation of SBMM. Or how the MM prioritises connection or chooses servers.
But everyone should like TrueSkill2. It’s just cold hearted maths.
The “algorithm” doesn’t need to be under handed.
If you have accurately ranked players and a match making system that fairly balances the two teams - the natural consequence of this is a 50:50 win rate.
It doesn’t have to do anything else.
It doesn’t need to analyse what your rank is or what you’ve done in your last handful of games. It doesn’t care.
And as an aside, I was quite surprised that when I found that if you take 8 players randomly selected from the population and sort them into two even teams you can pretty much get 90% of averaged teams within 25 MMR points. So you can imagine how easy it is for the MM to balance the teams when it is already choosing players closer in skill.
something happened to matchmaking today. Just lost 10 games in a row and rarely felt like winning.
maybe saturday is just a bad day to play halo because its probably when buddies decide to squad up. These where some of the sweatiest matches ever and i kept losing due to bad team chemistry and it was slayer too and still terrible
