This is a facile statement that is obviously untrue. If match results were randomized, top 500 play would resemble bottom 500 play. Is that seriously the claim you want to make?
1 Like
The best I’ve been able to gather is that he believes SR is tracked separately from MMR, so that SR is subject to Brownian movement if MMR is accurate, which it very much is, exceedingly so.
But it’s hard to pin down. Good luck.
1 Like
I’m just trying to pin down the [match results are effectively randomized] claim, because that is a very strong claim to make with significant ramifications if true. And I suppose I have seen that claim made before, but I don’t think it’s at all a common understanding of what is happening on the ladder, and I’m a bit surprised to see it being made so flippantly by someone who claims to have given the matter serious thought.
1 Like
That is not a claim that I’m making. We have heard in developers statements that MMR only applies from Bronze to Platinum, so it does not apply to Top 500 matches.
That’s not something I’ve ever seen. There are differences in how the system treats players above diamond. But nothing I’ve ever seen suggests that MMR is somehow removed for top 500 players. That wouldn’t even make sense. I assume this is one of those things that you have no citation for?
1 Like
This has been confirmed before and I don’t have it at hand (I don’t have linking privileges on this forum anyway). I refer you to WyomingMyst’s compendium of developer posts.
For those who don’t know:
“Brownian motion, or pedesis, is the random motion of particles suspended in a medium. This pattern of motion typically consists of random fluctuations in a particle’s position inside a fluid sub-domain, followed by a relocation to another sub-domain.” – Wikipedia
I know about this from Hank Green’s “Journey to the Micro Cosmos” channel on YouTube. It’s cool that Brownian motion can be observed on the same scale as the lives of many microorganisms, causing crystals and other particles to visibly dance in the transparent structures of their small bodies. I appreciate the reference, because randomness is a crucial aspect of the concepts we are dealing with. This is part of the discussion of MMR and algorithmic handicapping for sure, give me a little time to think about and phrase how it relates.
No. This has not been confirmed before. You cannot suggest something as nonsensical as saying that MMR is not used above plat and simply expect that claim to be accepted. When you make outlandish claims you must be able to support them better than simply saying, “this has been confirmed.”
1 Like
I just mean a random walk, man. I know it’s technically for particle movement, but we chemists like to use the term for every random walk, even if it’s not technically accurate. Apologies for the colloquialism.
2 Likes
For those who don’t know:
“Brownian motion, or pedesis, is the random motion of particles suspended in a medium. This pattern of motion typically consists of random fluctuations in a particle’s position inside a fluid sub-domain, followed by a relocation to another sub-domain.” – Wikipedia
On reflection I find that it does relate as an invocation of concept, but not in the way you are suggesting in reference to SR, more in reference to MMR and what that does.
Brownian motion is a (seemingly) random phenomenon that occurs in the physical world. Match Making Rating is a systematic phenomenon that occurs in the digital world. You could better think of algorithmic handicapping as the antithesis of nature and true randomness, with regards to matchmaking. If players are tiny particles bumping around inside of a cell membrane every which way, then the Matchmaker is a miniaturized surgical array of superconducting magnets making sure that things go one way or the other.
SR and MMR are not the same thing. SR is rank. MMR is a component of algorithmic handicapping. And though Brownian motion is not truly analogous to anything we see in these systems, you might say that MMR turns match results into Brownian motion. Because algorithmic handicapping makes the results intrinsically unrelated to the skill of participating players. Match results are uncoupled from player skill when their difference of skill – which is measure by MMR – is used as the basis of matchmaking.
This, if it is based on anything, is a severe misapprehension of what pushing for 50/50 matches means and how it relates to sorting algorithms. There is no attempt to ensure that matches go one way or another.
What does happen is that the matchmaker pushes for matches that it thinks will have 50/50 outcomes based on it’s current understanding of the skills of the players on the ladder. The reason it does this is to maximize any given player’s ability to impact the match, thereby ensuring that we learn as much as possible about the relative skills of the players on the ladder by making that match.
Otherwise we would make a different match.
1 Like
The analogy is imperfect and I’ve made it overblown, but I’m not the one who started it.
Wrong. Read developer statements about matchmaking and MMR, they talk openly and unapologetically about the engineering.
Yes! Do you not find this profoundly problematic in ranked competition? How can anyone accept this fact?!
But in so doing the matchmaker minimizes any given player’s ability to win matches with their higher level of relative skill. So how can SR gain/loss be tied to the win/lose result of the match at all? Shouldn’t SR gain/loss be completely dictated by PBSR, so that players could gain SR even if they lose a match? The system is completely inconsistent and arbitrary.
Also consider, the skill of players is supposed to be defined by Skill Rating (SR), so why is matchmaking based on Match Making Rating (MMR)?
I’ve long ago stopped trying to tell him how it works. The most I’ll do is try to get the concepts that I believe he’s missing into his head, but that’s not guarantee that he’ll apply them to this situation. I’m more interested in the details of his conception. I suspect that trying to pin it down is like trying to pin down half of a balloon. If you read my stories above, you’ll see where I think his mental errors are.
I really do think that you understand how MMR works, but for some reason you don’t want to speak plainly. I also don’t think you’re seriously asking a question here, mostly because how the system could work is repeatedly explained, yet explained away by you on the fact that there is no actual proof of anything.
I mean, explain it like I’m a 5yo please? Do you not think the problem is that MMR creates matches that have 50/50 outcomes, and that this is a problem because SR (the only thing that matters) is tied to a win or loss, so that wins and losses are completely random and thus, SR movement which goes up on a win and down on a loss, must also be completely random?
Like, I keep hearing you say this, but then you seem to deny it every time I try to spell it out. Where exactly am I wrong (on what you think)?
1 Like
No. I do not find this problematic at all, because I understand how the underlying math works. I wrote a long post on this in another thread, but the short version is that if we match more randomly we reduce the chances that any given replacement player would affect the outcome of the match.
If, instead, we match for a 50/50 outcome, we give any given hypothetical replacement player the highest chance possible to affect the outcome of the match. And since our 50/50 match was made based on our current understanding of the relative skills of the players involved in that match, we are most able to test the hypothesis that our current understanding of the skills of the players involved in that match is, in fact, wrong- thus allowing us to suss out fine differences in the skills of the players involved.
Systems that involve only a single skill rating that is equal to the current ladder rating are weaker than systems that involve both a hidden skill rating and a ladder ranking. The hidden skill rating is better able to measure player skill than the blunt force of wins/losses is able to capture on its own. That’s why we make matches with it rather than SR. If we are wrong, however- if our MMR is wrong and therefore our current best understanding of the skill of the player is wrong, our 50/50 matches will over time show that better than more random matches would.
These systems work in conjunction and they are better than the commonly proposed alternatives. I’ve demonstrated this numerous times and in various ways throughout these threads. No one has even attempted to demonstrate the opposite. And the claims that are made in opposition to the current system of hidden MMR and 50/50 matchmaking and performance based SR acceleration are often contradictory and poorly thought out. They do not even cohere with themselves.
1 Like
You’re the one who’s missing concepts, because you keep trying to compare MMR with SR and a billion other systems that it’s not. SR and MMR are not the same thing. Get that through your head.
I’m never going to get it through my head until you tell me what you mean. I know you think this, but I can’t understand what alternative you believe exists.
In my attempts to describe what you mean, SR and MMR aren’t the same thing.
1 Like
My god it’s not randomized it’s basic statistics that if you are playing with a bunch of people at your skill level you will win half of the time. Pay attention more in school then thinking about your rank in a video game 24/7
1 Like
I mean…I don’t know what he expects or what he actually thinks. This does seem elementary, but I have a hard time believing someone can be this bad at math. Unless he comes out and teaches us rather than just telling us we’re wrong, I don’t know what else to think.
1 Like
I mean he’s just jumping through mental gymnastics to try and say that He wants to climb without getting better. Which again never makes sense because you would think these people would be happy playing against people of their skill level than getting rolled. But no these people still can’t grasp that you don’t get rewarded by ranking up with playtime which is what blows my mind the most.
1 Like
This was THE APPEAL of OW when it first came out. It was the most amazing thing.
At it turns out, though, it is really bad for retention, winning only half your games. While it allows everyone to play, it makes all the games hard. When you combine that with teams of 6 individuals with varying attitudes, you get the chaos that these guys are, in fact, experiencing.
Conspiracy theories arise out of people trying to make sense out of chaos. Often, it’s easier to believe that a system is maliciously designed than it is to understand how the same experience is a result of a thousand small conditions interacting in complicated ways.
Conspiracy theorists stick around because they like being on an intellectual par with others, and even feeling superior at times. They’re not dumb, really, but they’re not taking a truly objective view. The more you engage in argument, the more “on par” they are able to feel. The argument becomes the point, rather than fact finding, as it is for the rest of us.
It’s important to distinguish the theory from the theorist. You can defeat the first by education. People who are seemingly willing to accept the simple answer are reasonable. We all do it, you, me, everyone believes in some conspiracy theories even if you don’t realize it.
You cannot educate the theorist. They’re working on a whole different set of assumptions. They must do educate themselves, which you can encourage by asking them to explain, in detail, how things work.
Insulting them will never work, because that’s what they want. Insults are, in fact, a “win” because it’s the admission that you cannot defeat their position on facts, which of course you can’t do because they deny the same reality you accept.
The only thing that will get a conspiracy theorist to stop believing their own BS is to get them to stop not believing, rather than start believing the truth. It’s not enough to just accept the chaos, they need to make an affirmative statement about the theory, at which point the same mechanisms that drive the conspiracy theorist will work to drive them to reality.
It’s much easier to doubt lies than truths, but it’s easiest to doubt everything.
https://www.theatlantic.com/ideas/archive/2021/03/the-conspiracy-theorists-problem-isnt-what-they-believe/618285/
2 Likes
I believe that has been his entire motivation from the get-go, he really thinks he was a GM player whilst having touched plat like once and his last recorded SR is like 1769 (im pretty sure anyone stuck in gold for 600+ levels isnt GM). So this is all likely the result of either a massive ego or a god complex.
1 Like