So SBMM is designed to decide when you lose…?

Games Halo Infinite
630

The only difference between the formula in Halo 5 and the one I was hoping to see in Infinite is how it’d allow people’s MMR to update first (post-match) before adjusting their CSR.

It’d simply remove the personal performance skill-lag that existed. I’m not sure why you’re referencing KPM and saying you can’t have it adjusting CSR & MMR. KPM and DPM are simply components of individual MMR adjustments within a team’s net MMR adjustment. They have no affect on CSR other then how they impact a person’s MMR.

631

It would be tricky.

For example, in an upset, your MMR goes up and theirs goes down. So the difference is lower. And this would change up your CSR calculation. Especially if the MMR is relatively volatile (which seems to be the case).

If they reduced the scale of the CSR to a point that it doesn’t change per game (like when it was 1-50) then the lag doesn’t matter.

632

Most of the time it seems like those with better KPM just have the better connection. So in theory, the richer you are (able to afford these luxuries) the better your KPM will be. Skill is relative in my humble opinion. Switch places with someone with desync and see how fast your KPM is then.

633

To some extent yes. And it’s a sad fact of life.

It’s probably not as bad as it used to be. There is a big benefit jumping from 30fps to 60fps. Not so much 60 to 120fps.

And yes. Better monitors can help. Fancier controllers. Comfier chairs etc. It’s a problem console gamers have been sheltered from up until now.

And pings are often as much luck with location as they are with cash.

True. The difficult part is putting an actual number on the difference.

In most cases we’re talking about what? One, maybe two tiers?

And really, it’s the “comparing” of skills that is relative. Improving your rank is still a rewarding pursuit.

If it’s any consolation - it seems to be people with better connections that are complaining about desync the most.

1 Like
634

In theory yes… in application I’m not sure this is implemented well… I love the gameplay, I personally don’t agree with the current ranking system… not sure if this is the same as sbmm however (edit… I love the gameplay a small but important maneuver I think needs to happen is to be able to repulse while sliding.)

635

Me too!

I’m all for the TrueSkill2 / MMR system. I think that works well.

Not so fond of the current way they do the CSR.

It sucks.

The SBMM is the Match Making system using your MMR.

I think it was working well before the population dropped off. The current set up of the teams, while “fair”, isn’t particularly “fun”. But I’m not sure what else they can do at the moment.

Do you mean to extend the slide? Or to boost out of it?

Interesting proposition… care to elaborate?

636

I hold right trigger to slide, when I let it go to repulse, it doesn’t work until my slide is complete. Then after I’ve pressed the repulsor two or three times it eventually works after I’ve spazzed hitting it three times. If somehow I could slide and while sliding repulse my opponent that would be what I’m attempting. Make sense?

1 Like
637

343 is prioritizing flashy animations over fluid, functional gameplay. Like how it takes a full 4 seconds just to get camo or overshield. Or how it takes a full 3 seconds just shoot the rockets unless you YY. Unlike Halo 5, which has the tightest and best controls in Series. Everything that they got right with H5 was scrapped completely. For no reason.
We have scientific evidence of Alzheimer’s and how a person can try to prevent it. But unless you’re living a healthy lifestyle, you go backwards mentally as you age. That’s life. Why does 343 go backwards with their design though? Why does 343 do the things that they do? I just don’t have the answers, man. You can’t even say it’s greed because even the store is a massive failure. Maybe 343 is just on a higher plane of existence than we are. Nobody really knows what goes the minds working at 343.

638

The only ‘fair’ way of matchmaking is open lobbies. Balance the teams based on the 8 random players that are pulled in. That is the only way to fairly represent where a player is in the grand scheme of things

639

100% true.

-Adding more characters so I can post this :slight_smile:

1 Like
640

Allowing MMR adjustments to occur first, so a CSR adjustment could include the most recent MMRs instead of relying on the pre-match MMRs very well could be tricky. And maybe they were simply unable to figure out how to achieve it without causing all sorts of other issues, but it’d be quite disappointing from my perspective if they didn’t even try or attempt to address the issues it may have caused.

In an upset, the winning team’s MMR is guaranteed to go up. But to be clear, a specific individual on the winning team isn’t necessarily guaranteed to see their MMR go up.

The chances of a player’s MMR going up is much greater when winning. And significantly so when winning unexpectedly. But, it still remains contingent on personal performance (via KPM & DPM) relative to a player’s expectation… and those of their teammates. This is because the team’s net MMR adjustment is dispersed amongst the team members (in a zero-sum manner) per the individual performance expectations.

The CSR formula being changed to incorporate post-match MMR reflections is the very point I’m after, but remember that the CSR formula would encapsulate more than just the MMR skill loss of an opposing team within an upset scenario.

Are you suggesting that they should revert to a visual rank system for the CSR that doesn’t necessarily represent an accurate representation of skill or reasonably adjust to a players skill over time?

1 Like
641

Number ranks are a great indicator. I truly believe that Halo 2 had the best ranking system because players were able to level up to higher levels and then continue to battle it out. Sure, there will always be one-sided matches, but where is the fun and push for me after I hit 1500+ Onyx rating. There is no change at 1600+ or 1700+ or anything else. At least in Halo 5 there was Champion for the top 200? players. Halo Infinite should at bare minimum have that too in the ranked playlists.

642

I was quite surprised when I crunched the numbers… you can actually balance teams quite well from 8 random players.

Just under 90% of games would be within an average of 25 MMR (on team average).

It would still be awkward with the Onyx player carrying Gold players vs an average opponent team of mid Platinum.

Probably OK for Social.

But then again, you can simply narrow the MMR range and make a better experience for everyone in the match.

And the whole matching of random completely breaks down if you have two sweaty Onxy players chilling together. You can’t really make matches with six other random players. Over half the games will have an average MMR gap of 250 or more.

Yuck.

643

I don’t have a problem with the divisions. Bronze to Onyx. They are actually standard deviations of the population curve… so you can estimate the numbers in each one.

And a gap of one division (300 MMR points) should give you a 75:25 win ratio.

But yes, I believe a smaller scale for the numbers would be better. No necessarily 1 to 50 like Halo 2. But the 1 to 1800+ doesn’t work.

It was a great start. And evolved into TrueSkill and then TrueSkill2. But you wouldn’t want to “go back” now that it has been tweaked to be faster and more accurate.

I’m not sure how it’s different now?

You are, hopefully, playing better players?

But yes, I agree, the CSR grind loses context. Better to have tiers in Onxy. Maybe they could have 1 - 10 and use the symbol ranks they had in Halo 2.

Oh God no. Horrible concept. Completely inaccurate.

What I would prefer is a rank points system. Every time there is a balanced game of Onyx vs Onyx and an acceptable ping then the game is for “championship” points. You then accumulate points through the Season - with rewards etc.

644

Probably too tricky.

You can’t have KPM change both MMR and CSR. It effectively double dips in the next game when MMR changes your CSR.

And I’m not sure they really want to put KPM front and centre to CSR changes.

It’s only a weighting. And it’s relative to your performance vs opposition MMR. You can go positive, with an impressive KD, and still fall short of expected KPM. That will just confuse people further.

We’ve already seen people here on Waypoint struggle with the concept of KPM vs KD.

I’m sure they have tried. But KPM isn’t an easy concept to sell to the masses.

Maybe they could start by introducing a medal for finishing the game with more than expected kills and less than expected deaths. That would at least introduce people to the concept.

I can see what you are after. And your intentions are good.

But I think it all falls apart when your CSR is above your MMR. You really don’t want your CSR going up too much in this case. You are only going to lose it all on the next loss. And if you introduce KPM to CSR they are going to have to either give you even more points that are only temporary - or inconsistently apply them depending on your MMR.

I think it would just confuse and aggravate people more.

It would work beautifully post placement. Your CSR would rush to meet your MMR faster on the back of good personal performance. But once you reach your skill ceiling and the grind becomes real - I think it would create too many problems.

I’m suggesting that the CSR scale isn’t a particular precise representation of your skill.

It’s supposed to be an average (or at least smoothing) or your MMR. It looks like the volatility of the MMR is left to fluctuate by 50 or more points a match (from Menke’s various MMR graphs).

At the moment players are accumulating and losing CSR points that are not reflecting game performance or change in skill. 10-15 points up or down has no contextual meaning. It’s just a visual rank oscillation.

But people are investing way too much emotional energy in these CSR blips. Both in terms of time and effort to “earn” them (including toxic play) and getting upset with the system when they lose them.

We know that your overall average rank doesn’t change on a game to game basis. For most people it doesn’t change day to day, or even week to week. So why are we making it look like your rank actually changed on the basis of that game?

If you had a smaller scale. Say 1 to 100. You can just get on with playing. Your MMR can jump around in the background as it needs to. And your CSR can fluctuate at a less stressful pace - drifting up or down over the course of ten or more games.