SBMM Has got to go

No, I just think that there is no merit to any argument that goes something like “This implementation of True Skill 2 is good because it improves player retention”

If I develop a drug that, in clinical trials, cures a disease, but then that drug does not in fact cure any disease when it is released into the general public, the drug does not actually accomplish its intended purpose, rather or not other external factors are the primary reason for the failure.

Again, the average player you play will wind up being smack dab middle 50%. You will get to play against bronze and silver players as much as against diamond and onyx players. You are attempting to argue against particular potential experiences that a person might have and might react badly to, but you aren’t addressing the fundamental question in this debate - Why is variation in experience bad in unranked game modes?

Ranked game modes have historically (pre- Halo 5) served to be the “control” for skill based matchmaking. If you are bronze 1 and want competitive 50/50 matches, something approximating that is available to you.

I think it’s a pretty good case (one of several) for abandoning the current one-global-MMR-to-rule-them-all-in-every-list SBMM.

It’s obvious you think it’s fun, but it also appears that you cannot comprehend why others do not think it’s fun.

There is a fundamental philosophical assumption behind your argument - every player, regardless of their skill, should have basically the same kind of experience in the game, at all times. I don’t agree that is fun at all.

I believe this assumption fundamentally removes a significant piece of the incentive to get better at the game, because the game actively tries to not reward players at the higher end of the skill curve in terms of their gameplay experience. I’d argue that in some cases it actively punishes players for being good, such as when it decides to put three absolutely awful teammates on your team to make the game’s final score “close”.

I do understand that there are times when 50% winrates and nail biter games are fun. That’s always been the goal of ranked playlists. I do not want to see that changed in any meaningful way (although I do think it would be much better if it were to match based on CSR instead of a hidden global MMR affected by everything).

However, I believe that providing a much greater variety of experience in the social, unranked playlists - one that better approximates a “random” sampling of the population - provides something of great value and fun for the game, that has been sorely missing since Halo 4.

I also don’t agree with the assumption that it’s bad for a lesser player to get stomped by a better player. If theater mode worked correctly, that provides a really great learning opportunity for the worse player that they would never have if the matchmaking is as it currently is. For example, on launch day, I got to play against Mikwen one game in the ranked list, because there hadn’t been enough time for him to separate out from the rest of the pack climbing from Diamond to Onyx. It was absolutely fascinating, even though I got absolutely dumped on for the entire game and felt like I was a helpless Bronze. But even with the broken state of theater, I got to go back and watch what he did, and I got to learn and improve as a result.

I think that the current system only benefits players who cannot emotionally handle getting dumped on by a much superior player, and who would consider quitting the game if that ever happened.

1 Like

Ok. Except that the people who are saying it are the ones who actually have the data at their fingertips.

No. You would go back over the clinical trial - examining the thousands, if not millions, of data points, and look for what correlations (and their strengths) that you missed.

If you have something that clearly improves player retention, even to a relatively small extent, you don’t drop it. That would just be making the problem worse?

That is true. They will definitely have the “fairest” experience.

Variation is bad if people have different experiences.

We can clearly show that those to the left of the curve are having the opposite experience to the right of the curve.

That’s not fair.

And as someone pretty much in the middle of the curve - I don’t want to be stomped… or to be stomping. What is the point of that?

Not sure how?

Games like Fiesta actually increase the skill gap.

I have no chance taking down Mint Blitz if he has a rocket launcher and I have a pistol. Zero. And sadly, if the tables were turned and I had the rocket launcher, I doubt the outcome would be much different.

And do we have any proof of this “one global MMR to rule them all”. I see it discussed a lot - but it kind of goes against how TrueSkill2 would normally be implemented. And it certainly wasn’t like that in Halo 5.

Besides, I’ve been keeping a close eye on how ranked works each Wednesday (after spending Tuesday nights horsing around with mates in BTM and other social games). I can’t say I’ve noticed any difference. And I’ve started a spreadsheet for the average MMR of my opponents - and with the small sample size I have to date it looks pretty consistent.

Fun is in the eye of the beholder.

I can see why it would be fun to the top third, running around and showing off. Chalking up multi-kills. It’s not my cup of tea - but I can see the attraction. And maybe I’m just jealous.

I can kind of see why it could be fun for the middlers. Again, I wouldn’t like the stompings (in either direction) - but overall, OK.

But I’m yet to see how it would be fun for the bottom third. Smashing after smashing after smashing. All we need is the calls of “get gud or get out”.

Oh wait… you did. Just much more eloquently than most.

One of things I found interesting while doing a bit of googling about SBMM was this quote from a developer…

It seems to fit my personal bias in stereotyping those who are anti-SBMM as OG Halo fans who worship H3, hate advanced movement mechanics, are infuriated by coatings, and stubbornly refuse to spend any money on microtransactions.

Maybe it’s not so much that SBMM improves player retention… but that SBMM improves player spending. :slight_smile:

And yes, I’m being a bit tongue in cheek here.

There is actually a super cool GDC talk it is only 27 minutes and actually explains how truSkill2 works look it up on YouTube.

That may be but:

I’m simply showing you that in fact the “good players” don’t enjoy SBMM.
But his video gives good rounded reasoning why it is in everyones interest to not want SBMM for social playlists.

youtube.com/watch?v=F9qMkK3wLhA

If you watch the video he tells you about how the original creation of the “TrueSkill SBMM” is not the result that the developer wanted to make. But was forced into it.

I realize what the title of this thread is. But just because I disagree with the current implementation of SBMM, doesn’t particularly mean I want it gone entirely.
We are simply saying that Social vs Ranked makes little difference. What is the point of having Ranked or Social if they play the same. In fact I find social sweatier.

I don’t disagree with team balancing as SBMM. However I want them to balance a team with a larger pool.
Lets say you scale as a level 35/50 in MMR. I want the game to compose matches that vary from 25 → 45 instead of 33 ->37.
I think that the teams still need balancing however. So if you are in a fireteam of 4 players all with MMR > 45/50 then yes you should be matched with people of similar skill… However if you are going in solo I want to be the player that potentially carries one game… then gets carried the next… Variety!!

But no I’m not suggesting that the game just match 4x 50 MMR players with 4x 15 MMR players “randomly”. (No one has said they want random).

this is what I believe most people are trying to explain here. We want balanced teams… not narrow skill margins that equal constant sweat.
This means that ol mate minty will be carrying most games due to being high skill bracket. But he will actually find games locally with a team MMR balance that looks like.
RED: 50, 33, 27,20
BLUE: 45, 43, 39. 33

Anyways… regardless this is what MCC is doing right now. And that is why I will continue to play MCC over infinite or halo 5 (as well as some other reasons relating to low population pools in Aus).

I’m a good player and I like sbmm.

Can understand why someone who makes a career out of clipping players that don’t shoot at them doesn’t like sbmm.

1 Like

I pose my question to you then. How would you describe the difference of social vs ranked?
Other than simply saying you get ranked in ranked. Do you agree that social and ranked basically play the same? When social in it’s definition is to be social… casual… relaxed.

Like playing leagued sports vs backyard sports… You can very easily get stomped playing sports socially… and still have fun! It’s also how you get better.

If you are getting stomped hard in social then go play ranked… It’s what I did when I sucked at Halo 2.

I propose this is 99.9% your mind set going in.

Combined with the fact that there is little to differentiate Ranked vs Social in terms of structure. It’s essentially the same maps, lay outs, and rules. It’s no wonder games mimic rank when the heat is on.

Personally I think this would just makes things worse.

We know that wide range of ranks tends to break match making in ranked. And can even be manipulated by carefully constructed squads to milk CSR.

And you automatically think it makes it easier for the lower ranked players to compete. No, the are just appetisers for the good players on their way to fighting each other.

You get away with larger ranges for BTB. But in a 4v4 it starts to break down.

Which is what we all want. Ranked matched on a single player. Social matched on teams (but for reasons above it has to be moderate ranges).

Personally I would revamp Social entirely. Move away from mimicking what is happening in Ranked. Different weapons. Different game types. Some degree of handicapping to level the playing field.

Something different.

Again. It’s a mindset. I’m not very good but I sweat my -Yoink!- off in ranked. Try to get every last inch out of my (limited) ability.

In social I muck around. Try new weapons. Go for spectacular kills. The fun stuff.

I don’t really care about the result.

But some people just can’t let go. If their team isn’t winning they feel compelled to switch into ranked mode and carry their team hard (and then proceed to whinge about it for hours on end).

Sweaty. Chill. It’s a frame of mind.

Fair enough. I obviously can’t counter you saying that is your state of mind. However, in terms of it being “functional” see halo MCC. Halo MCC operates on this principle any I couldn’t be happier about it. For one I can actually get low ping games. (which I admit is a huge reason why I am passionate about social being social)

In terms of being appetizers… I agree that if you are like MMR 3/50 then yeah that’s a tough gig. That was me in Halo 2. I was still learning how to use two sticks on a controller having just moved on from n64. But that is why I played ranked. So that I would be matched tighter.

As someone who lives in a low pop area and therefore has shared a lobby with mint blitz many times and has gotten stomped, I would still choose that variety any day over game after game that plays the exact same. I suppose with the right frame of mind my morning commute could be enjoyable.

1 Like

In ranked, I only play when I’m fresh, wear turtle beach headphones, listen to footsteps, callout and ping a bunch, focus hard on what I’m doing and the team. I always open with meta strategies and focus hard on map control. The games are often extremely tense and extremely close. I’m always thinking ahead about next steps. I’m in Onyx and have a 61% win ratio in Crossplay.

In social, I play whenever, sometimes very tired, listen to music or documentaries, never use my headset. Sometimes I will open with a meta strategy but often I’ll just do whatever I feel like doing. Regularly I’ll be going for challenges so instead of contesting rockets I’m running straight to the needler.

My win ratio on social is as follows:

Swat - 59%
Slayer - 63%
Quickplay - 59%
BTB - 51%
Fiesta - 64%

Most the time I’m playing social I’m doing challenges (sometimes that means being an absolute filthy tryhard and sometimes that means doing awfully as I try to do a basic action). The crux of it is I don’t have to put much effort into social compared to ranked. Usually ranked is highly intense, whereas social is often not. Some games are, but most games are fairly casual. I have great win ratios in social despite taking a very casual approach to them most of the time. It’s not like I’m trying to lose but the level of intensity is much lower.

BTB is one of my favourite modes and my lowest win ratio which is still 51%. I spend most of my time playing BTB fooling around with teammates and vehicles. I use a lot more of the Sandbox and particularly love getting kills with the fusion coils and turret on Fragmentation. Sometimes I turn it on, usually if there are other players on the enemy team beating up my teammates a little too much. I tend to only do really well if I am doing challenges.

It’s just clear as day to me that whether I’m top of the leaderboard or near the bottom, the game is still fun for me. I’m just playing how I want and freely. Compare it to ranks where I have to be 100% on at all times or we lose most the time. The games are so close so often it’s just a fact you can’t ease up. In social half the games I don’t turn it on at all, or I only need to turn it on in the final quarter to take the win back.

1 Like

Interesting. Reading your response, I can’t say I understand your point of view that you want SBMM to remain strict in social. Everything you just said suggests that you see social as casual… sometimes sweaty sometimes a breeze. Sometimes doing challenges and sometimes just mucking around.

Does it bother you that your MMR is shared across social and ranked? ie if you jump on and warm up in social then switch to ranked that your win streak in social affects your ranked games?

I don’t believe it should be strict in social. I’m saying it isn’t that strict right now. Any looser though and I might take issue with it if I’m constantly playing much worse players. The level now is pretty good. I still have to play fairly well each game but it’s nothing like ranked.

To put it lightly, your MMR number isn’t the same for every playlist. Being 1,700 MMR in ranked won’t necessarily be putting you in 1,700 MMR games for Grifball. The correlation between playlists is driven by player data. Most playlists we have now likely have a fairly high correlation but still doesn’t mean it will be exactly the same for every playlist and that games will have a major effect on the outcomes. I play a bunch of BTB (my worst playlist) yet that didn’t noticeably slow me down from climbing a few hundred CSR points.

Even if it did, I can’t say I care. I don’t care what rank I am provided I’m getting good close games in ranked. I’d rather be 1600 Onyx having 4-3, and 3-2 CTF games all the time than 2000 Onyx having 4-0 games. I’m not in ranked to get a higher rank, I’m there because the games are more intense and have higher stakes. Of course I like ranking up, but it’s not my primary purpose for playing ranked.

Actually MMR is your hidden matchmaking ranking. You can’t see what it is. But it is how the game ranks you when matching. (regardless of playlist).
Your playlist CSR ranking is the playlist specific rank. It doesn’t correlate to MMR which is why you can see a gold player who did badly when being placed in the first 10 games, and is slowly getting his CSR raised whilst being matched with Diamond players.

In rank playlists everyone agrees that tight competitive games is the goal.

I haven’t actually played infinite for a while… but it was tight on release. Maybe it’s more relaxed now. But to this day I cannot match in halo 5 because it was too strict. I simply could not find a low ping match. I would find a match when I played with a friend who is no good at the game. Or once I made a smurf account and I got plenty of games… but it only lasted a little whilst until the MMR was too high again. Halo 5 was definitely too strict.

Your MMR won’t necessarily be the same across every playlist. There is an offset skill distribution per game mode based on level of correlation which is determined from player data.

They did an update and I’m certain it got made looser after. The game was super strict at launch in social. Try it out.

The skill gaps between individual players (like you ask for here) are already very big. Just look at my matches from yesterdat and tuesday:
Yesterday:
12-4 Quick Play, Oddball
18-3 Ranked, Slayer
23-5 Ranked, KOTH
37-8 Ranked, Strongholds
7-8 Ranked, Strongholds (it was 2v4 though, so i had was in 1v3’s or 1v4’s every time)
17-7 Ranked, Slayer

Tuesday:
13-2 KOTH
9-7 KOTH
10-4 Fiesta
21-1 Team Slayer
9-3 Quick Play, Attrition
21-2 Quick Play, CTF
23-8 Quick Play, KOTH
2-2 Quick Play, KOTH
18-12 Quick Play, CTF

Does that like a very strict SBMM with very little gap between players? It is clearly having big skill gaps, even in ranked.

A lot of people disagree with you here, they don’t want to be carried by way better players, because they are just outskilled. Often they just quit those matches. Apart of the challenge system and crashes, this is where the quitproblem comes from. People just hate being in matches where they don’t stand a chance. It doesn’t matter if they win or not, they want to be able to play and not just be cannon fodder. On top of that you also hear the better skilled players complain that they have to heavy carry to be able to win matches and don’t want to do that. The problem is not that the skill gaps between individual players is to small, but to big.

You may not be, but there have been a lot that are (this is only the millionth topic)

You might be, but i have heared a lot of entitled players ranting that they don’t want close matches.

This is deliberately framing the question, since you put it so that you won’t allow the main and most important answer.
But other (less important) differences are:

  • different starting weapons (BR in ranked, AR/SK in social)
  • weapon spawns (the spawns in the racks are not always the same as in social)
  • radar (not in ranked, available in social)
  • gamemodes (limited gamemodes in ranked)
  • melee handling (BR 2 shots+melee in ranked, 3 shots + melee in social)

So yeah, there is a clear difference. The reason why you don’t ‘feel’ that difference is because of your own mindset.

No they are not. The ‘stress’ of having to win makes it play different already. Beside that the starting weapons are also different (BR <-> AR/SK) and those weapons also make you play differently. And no, social does not mean casual. Casual is just a mindset. You can play casual in ranked and still dominate, it depends on how much ‘stress’ you put in performing. The reason many complain about ‘sweating’ in ranked is because it’s their own mindset that is in the way. They complain about not being able to play worse while performing just as well (k/d wise) or even better and win the match. Ofcourse if you play worse you perform worse. The whole point about social is that that is okay, because you don’t have to win. Yet they just can get that in their mind and therefore complain about having to ‘sweat’. They just keep the same mindset as in ranked.

You can also get stomped playing sports on the highest level. Do you remember the stomping Barcelona got in the Champions League semi-finals against Bayern München (8-2)? But also here your comparison don’t fly, since the top teams play in the top competitions and not in between the smaller teams. That too is a form of SBMM. Just like in tennis only the best player will compete at Wimbledon and the lesser players have their own tournaments. Etc. etc.

Not asked to me, but i would say that MMR shouldn’t be shared between ranked and social yes. In fact: i think MMR (except for placement matches) shouldn’t even be in ranked. Ranked should match you based on your CSR.

I think it would be hard for me to judge. I stopped playing infinite because of the 150 → 250ms games.

If they have indeed broadened the SBMM range then maybe I should see if I can get local games. But it possibly isn’t a valid indicator that the SBMM truly is wider because there are probably other reasons why I might continue to get 150+ms games.

It appears that Halo Infinite may have changed since release. Perhaps I would agree with you next time I play it.

The reason I don’t “feel” the difference is because SBMM keeps matching me against americans with 150ms ping just to place me in close MMR matches… ranked or social. Connection should trump MMR in social. SBMM should fix skill gaps whilst prioritizing connection… or simple you will lose the international crowd across the pacific. Why do you think I haven’t played infinite?

Do you agree that Halo 5 had strict SBMM?

This might help me understand if we are on the same page.

Since you say ‘across the pacific’ i assume you are from Australia/New Zealand or East Asia? Problem there is that there are not enough players, so either the skill gaps must be enormous or you have to play on US servers.

I am from Europe and most of my matches are on the Amsterdam server where i have just 10-12ms. But from time to time i also am being put on US servers with 120ms.

For your question about halo 5: i haven’t played that one. I don’t have an xbox (i play through steam) and H5 is not available for pc.

Well Halo MCC enjoys a healthy australasian population. And it has a nice social setup with a varied SBMM… Admittingly match composer and server selection also contribute to this pleasant experience.

I will have to try infinite again… but the SBMM was the nail in the coffin for my region… thanks to 343 for not adding match composer, not adding server selection… blocking geofiltering… and then employing strict SBMM.