Activision Blizzard released a 25 page white page document with an A/B test from early 2024 where they kept loosening the constraints of SBMM and monitored retention and turns out everyone hated it.

nanoUFO@sh.itjust.worksmod to Games@sh.itjust.works – 51 points –
activision.com
16

I mostly play games that are so niche that the matchmaking simply consists of "whoever's available". But the idea that being matched against opponents at your skill level is somehow a negative might be the most bewildering discourse I've heard in a long time. Genuinely why?

That's not what the paper says. This is specifically COD games that this was tested with

The loosening the skill matchmaking found players leaving from the bottom and continuing as new players found themselves at the bottom. Higher skilled players liked this as they got treated as having lower skill as lower skilled players left.

Tightening it found higher skill players leaving due to longer queue times and having less lower skill players to beat on in their matches. Lower skilled players had higher retention due to being more likely to be matched with their peers.

In other words high skill players enjoy stomping noobs more than fighting each other. Noobs don't like being stomped.

It's not entire untrue to say "everybody hated it", but it also misses the point.

Yeah so matching players with opponents at their skill level is a good thing. Should've been obvious to anyone with half a brain, but somehow there was a whole saga where people were radically opposed to it for some reason. And it turns out they were full of shit.

All I ever wanted was good fights and for that short time where SBMM was actually what it said on the tin, I got them. But then somehow devs started getting it in their heads that what people really want is noobstomping, and while I'm sure some people want that, they can go fuck themselves. That shit made me want to play less because it was too easy all the time, and then when the cheating became too rampant to ignore I just stopped playing competitive shooters.

But then somehow devs started getting it in their heads that what people really want is noobstomping

What dev ever said that?

It's the eternal debate: Should you, as a parent let your kid "win" when playing games, or should you play fairly and crush them until they either give up or get skilled enough to actually beat you?

There are pros and cons to either solution and ultimately it depends on what the individual wants; the immediate satisfaction of a balanced experience, or the assurance that every win or loss was earned fair and square.

I don't play these types of games anymore, but as a teenager I played a lot of Battlefield and I went from noob who would get absolutely crushed every game, to good enough at some game modes that my presence in a 32 player lobby would be sufficient to tip the whole game in my favor and my team winrate was well over 50 %. That is a meaningful, long-term reward that does not quite compare to the modern approach where no matter how many hours you sink in honing your skill, you'll still only win about 50 % of the time. Yeah sure you have a fancier badge or whatever, but it doesn't feel like improvement.

Of course Activision makes a compelling argument that SBBM is overall better for the health of the playerbase. I do feel like we lost something though, and that it is another area in life where algorithms decide what our experience is going to be and smooth out any meaningful challenge.

Hard disagree. I neither like dunking on or being dunked on. The best games are when it’s a close match, and you know you played well but the other team did too.

The key issue that's hard to address is making a hard fought loss feel more valuable than any other loss and not worse than any other fight.

Some games a hard fought fight can look like rushing to the point, getting a kill and a trade and then spectating either the rest of the match or the 20s respawn timer before making the 30s run back to the point, rinse and repeat. This might mean you're "playing" for less that you 10% of the time you actually spend in the match.

That has more to do with game design than matchmaking, but yes, it’s hard to balance games so that dying has consequence but isn’t too un-fun.

It is, but it's critical to causing the goal of SBMM being desirable on both ends

I don't think that analogy makes sense. The parent and child would be two players with a massive skill gap between them. The point of matchmaking is that you don't match them against each other to begin with, rather than asking the parent to hold back.

Kudos to the paper's author for using the colorblind-friendly viridis color palette or similar in the graphs. Would like to see that kind of being the norm for graphs, unless there's a reason to use something else.

That's cool but I still want community servers. I don't care about SBMM.

Play a different first-person shooter that does have them?

Thinking back to Quake, there are some drawbacks, like the fact that FPS clients don't have a fantastic history of being hardened against malicious servers. I guess if you just play with people you know, it's not as risky.

I don't like playing without a SBMM system, when I was young and could pour hours on end into a game it was fun, but nowadays I would love to just play with others who can't dedicate half the day playing. My time is limited already, I don't see enjoyment spending it getting stomped by people who have more free time then me. Thats a BIG reason why I stopped playing COD and switched over to more casual style games

this is also likely why it seems you play a lot better when party'ing as a group that takes up half to most of a team, because the game has a harder time matching you against others as it can't balance the teams because it can't split the party

This is why I like this company.

Got a lotta reasons to hate them too, but I have reasons to like them