I don’t mean anything toward or against OP, just stating the difference.
Sorry. Asperger’s. Definitions are important to me 
1 Like
I didn’t make any claims purporting that which I don’t know to be truthful, or factual. You might have a dragon in your backyard, who am I to say you don’t?
As I said, neither you, nor I has any information on how the MMR functions, because the devs haven’t disclosed any such information.
And I say this respectfully, not to insult you.
Do the developers know? How do you know they know if we know nothing? How do you know we know nothing if what we know is nothing? But do we not know something then? Maybe nothing…
…my point is: These comments of yours are useless and lead to exactly nothing. You can use your time more efficiently in other ways. These comments are self-refuting. Claiming everything is equally true because you cannot disprove it is not how knowledge and truth works. The arbitrary is not knowledge, for it is arbitrary.
1 Like
Your argument seems to revolve around a player pool totally reliant on Blizzards Match making algo’s to get a group. Going back to vanilla WoW it’s been common knowledge that a pre-made group (one not assembled by a blizzard algorithms but one put together by a human who wants to organize their own group for the sake of winning) usually has an advantage over “Randoms” the same has been my experience in Overwatch. If the game did not have a bunch of utilities to help people assemble their own teams your argument may have approached semi-validity, but given you provided no real statistics to back your statement up, and didn’t demonstrate any kind of in-depth knowledge of how matchmaking works I honestly think your post sounds like a malformed rant.
1 Like
They know that we know that they know that we know
Dragons are real.
My comments are not useless, I’m simply stating neither of us knows that which we have not been privy to. Do you disagree with this notion? Perhaps you’d like to elaborate a bit, because I’m failing to understand how stating the truth of the matter shuts down any further discussion regarding the topic.
In fact I do. There have been multiple substantiated articles on how the Overwatch ELO system works. Just Google it (a great one is on towardsdatascience). Or use the forums:
Your claim that we know nothing about how the ELO system in Overwatch works is nonsense.
But here we run into the issues with Subjectivist reasoning. How do we know they didn’t change it since the time the articles were posted? How do we know they didn’t change it yesterday? How do we know the authors are not lying? Alas, we know nothing.
Edit: I have removed my last paragraph. In hindsight it was a tad wrong. Sorry if you were already typing a reply to it.
1 Like
That’s not what a professional gambler is, but it does meet the definition of a charlatan.
If you could consistently beat the odds for literally anything involving gambling you’d be milking the millions off of taking your own advice and not charging others a fee for such a nonesense “service” lmao. I’m better off taking lotto numbers from fortune cookies.
1 Like
I don’t know too much about statistics, but I do remember the class I sat in with my wife where her professor came in and proved this to be wrong ~ the thing about the coin flip is that you have the exact same odds no matter how many times you flip it. It does not increase or decrease, because it’s always the same every flip. Even if you get heads 100 times in a row, that doesn’t mean you’re more likely to get heads or tails on the next flip, it’s still, and always will be, a 50/50 chance of heads or tails.
Humans mentally assume that because they got heads 100 times in a row, surely the next one will be heads too, because it’s ‘logical’, but that’s just a thought trap we fall into because we like to assume there’s pattern to things.
And this is where context matters. Previous coin flips do not determine the outcome of the next coin flip. That is what you’re talking about. But it’s not what xx9 is talking about.
If you flip a coin 1.000.000 times, you will see a ~50% tails/heads ratio. And the more you flip it, the more it will converge to exactly 50%. That is what xx9 is talking about.
3 Likes
I agree and have been studying it in my own capacities as well as a programmer.
As I’ve mentioned before, there are two significant events that cemented that Overwatch is rigged.
1 - Scott Mercer mentions in this article that the matchmaker won’t form a match until a minimum winrate of 40% for a team is “fixed”. This concludes EVERY match has a 40%-60% chance of winning, minimum NO MATTER WHERE YOUR SR STANDS IN RELATION TO YOUR MMR. Masked under the guise of “fairness”.
https://us.forums.blizzard.com/en/overwatch/t/groups-and-matchmaking-in-overwatch/134776
2 - Jeff Kaplan mentions in this interview that MMR is what has always been used to match games, not SR.
https://www.youtube.com/watch?v=bn8-aWPvLwE
This means SR is completely irrelevant. Hypothetically, if you’re a GM and fall to gold. You will NOT be against golds, you’ll be against smurfs and other GM’s that fell to gold or you will be handicapped with even worse players on your team to make the match reach the 40%-60% predicted winrate and you will have only a MAXIMUM of 60% chance (predicted) of winning games down there.
You shouldn’t have to verse people you’d be versing at you PEAK just to CLIMB.
Anyways, y’all, keep poppin’ those blue pills.
Edit: I should also mention the inverse of this can be true, too. If you manage, somehow, to climb your SR ABOVE where your MMR rests, your presence will negatively affect the predicted winrate so you’ll likely be propped up with the better players. This partially explains both sides of the Overwatch is/isn’t rigged arguement. “Well my games are fine! Therefore OW isn’t rigged” - Boosted Mercy main whom’s SR is higher than their MMR.
3 Likes
Real talk, there was an article or book that I read in high school that one of my math teachers based an in-class exercise on where we had to make an educated guess on what set of numbers were faked from coinflip tests and cast a vote.
The real ones were never voted upon. The fake ones always were. (I read the article/book long before we did the exercise and everyone thought I was crazy, then the teacher said I was right).
Real experiments do not resemble heads-tails-heads-tails-in-that-order, ad-infinitum. In fact, the real ones NEVER look like that. It’s entirely reasonable to get 5-10 heads/losses in a row. That’s how math peeps know if numbers were faked.
Doesn’t mean the system is rigged, though.
Besides, claiming the system is rigged because its matchmaking tries to create fair games is a bit weird.
Hypothetically, if you’re a GM and fall to gold. You will NOT be against golds
You don’t seem to understand how it works. If a GM falls to Gold his MMR will not stay at GM levels. It will adjust to Gold as well and he will be matched accordingly.
Where can you prove that? This is based on the assumption that MMR moves at the same rate as SR. It likely does not.
Other’s have mentioned in the past that MMR is only calculated on wins although I have lost the source for that. I’m skeptical of that claim however.
More flips means that it’s more likely to converge to 50-50, but it’s more unlikely to be exactly 50-50 - if that makes sense.
Numerical examples:
10 coin flips yields 4 heads and 6 tails. Certainly possible and it yields 40% to 60%.
10000 coin flips yields 5027 heads and 4973 tails, which is 50.27% to 49.73%. This is closer to 50-50 but it’s still off by more flips (27 compared to 1). So while we’re converging to 50-50, it’s incredibly unlikely for us to exactly hit it exactly because, like you said, each flip is an independent 50-50 chance.
1 Like
Okay, but afaik Kaawumba is not a developer for the Overwatch team and had no involvement in coding the algorithm for the MMR. So his ‘theory’ on how it functions is as baseless and moot as someone making a claim to the contrary. For example:
Here we see Cuthbert making a claim in the same year as Kaawumba’s article that Competitive Overwatch games are handicapped by the MMR or ‘rigged’ as many have suggested. I cannot say Kaawumba nor Cuthbert are correct in regards to the issue being discussed or the claims made in their topics, because neither party fully understands or is privy to how the matchmaker functions because only the developers have access to this knowledge. This knowledge has not been made public, and probably never will.
The only thing people have to rely on is anecdotal experiences in game. That is, until Blizzard fully discloses their MMR algorithm. Which they won’t, because it is their proprietary property. And they have every right not to, just as players have every right to question things that affect them in game and to exercise caution in choosing to remain playing that game in the very near future.
So, back to the topic of invisible dragons in the backline, seriously, who tf am I to say you don’t? It’s mighty rude to dismiss someone’s claim without having all the facts, don’t you think?
Because that is how ELO systems work. You lose --> your MMR drops. Overwatch has the extra layer of SR above MMR, but MMR is primary.
Having a basic understanding of how ALL the systems of matchmaking implemented across all games, is generally considered is better evidence than saying “you aren’t the developer, you can’t know that”.
Even if I don’t know the exact method they made, if I know other methods can’t be implemented cleanly (and have read the white papers regarding matchmakers shared at stuff at GDC/online/GitHub readmes, etc), then it’s simply Occam’s Razor.
The most straightforward, most adopted version of a matchmaker is MORE than likely to be correct. It’ll be tweaked, so specifics might be off, but the methodology and underlying technology is correct. Therefore, my speculation would be sound.
There’s an entire logical fallacy based on this – the ignorance fallacy.
3 Likes
The thing is, why would they do that when the natural outcome of an elo system is that you get a 50/50 winrate?
Like, why would they give that friend of yours a loss streak in gold instead of just letting him climb to diamond? When he gets to diamond, all that’s going to happen is a completely natural loss streak because he’ll eventually up facing people better than he is.
I think there’s all kinds of issues with the matchmaker including odd streaks but I think it’s bad design/bugs, not intentional. Because I just don’t see a point to it.
It IS how ELO systems work but it would be foolish to assume that it moves at the same rate.
SR ONLY serves to limit the pool of which players to pick to rig the minimum predicted winrate to 40%. Even then, I’ve seen silvers in the same games as diamonds so perhaps it doesn’t even limit that anymore.
Real talk: Why is MMR hidden?
Because then quick play would be competitive.