If you’re unwilling to watch the dev update, then its clear you have no interest in truly determining the direction of the Development team, furthering my point that you only wish to serve your own agenda rather than help the community.
Most certainly I could do this (as I did above OMEGALUL) but rather than cherry pick, leave argument for misconstruing, or twisting the narrative, I’d rather you put in your own efforts to actually do something beneficial for the Overwatch commuity instead of spreading falsehoods.
Also, just for the sake of conversation, I think many people do not have a clear understanding of the difference between 50/50 odds and enforcing a 50/50 win rate.
These are entirely separate things!
A 50/50 odds to win is simply pooling players of equal skill to maximize the ODDS of winning, with 50% being the fairest possible odds.
To enforce a 50% winrate, you would need to actively pair a player into games with skewed odds. Such as a player on a losing streak on a team with odds above 50%. I think this what this topic is primarily attempting to question.
The odds of a game are just a mechanism by which we can gauge the match’s fairness.
Doing anything outside of that is essentially tampering.
the dev update won’t remove this, cause never admits this. They won’t neither really fight smurfing and eloboosting in a real way.
This handicapping MMking have been crafted cause they, already before the launch of Overwatch, have chose to allow completely smurfs in this game, despite not being a free to play (no new users flow in playerbase), this why:
A 1 time box sell makes stakeholders of a company happy for the short term
For the long term the playerbase will suffer, who will remain in the title are the most affectionates, and the new users instead trying ow for 14-20 euros, goes to apex (during this time they can move the devs on other products making more money than a long term project like LoL, who will suffer are just customers/athletes)
Then, blizz needs to don’t allow low elo players to to realize that even in the leagues where they were usually good they are annihilated by much better players with “strange new accounts”, how? Handicapping Matchmaking
That’s okay, and sorry for my own delayed reply. What a high-value contribution to the thread, and I am happy to see that you’re still adding more value to your tool.
I have been wondering, can you make a read-only version of the page, with a ‘Generate New Match’ button, or something to that effect? I believe the previous version of the tool allowed write permissions for anyone with the link, and this version was not editable so I was not able to use the tool.
Very well deserved, I appreciate your dedication to this subject.
Oh, then maybe the problem I had was in trying to use the tool in the same way as the previous one, i.e., editing the SR cell. I’ll try again! I still think button to generate new values would be helpful and users will be looking for something like that.
I’m going to try out the new functionality you mentioned. I think that illustrating group dynamics is a very good idea, illustrative of real world matchmaking conditions. To that effect, it might be good to add a cell/field for wait time, since high wait times are correlated groups and individuals that have disparate SR and MMR, relative to peers in their median rank.
Edit: I see now that the direct input fields are editable, but the SR Match Target and SR Match Result fields are no longer editable. Should they be? I suggest adding a new button for randomizing/generating new match example values.
I just fixed a decent bump in the calculation filter. It was allowing tiny values outside of the range of 1000 SR into the calculation, basically breaking things!
And, one solution I will do is get the tool up and running on github.io!
I can also convert the javascript in the sheet into python and put it on Google Colab. That way everyone can test the tool in their own space, see the code and how it works, and poke fun at me for basically using code as a brainstorming scratchpad!
As for a button, I can see if something like that will work. But the way a spreadsheet works is based on input values, of which are then passed into functions and calculated post-update. A button is more familiar to realtime interfaces, such as a webpage.
Edit: Whoa! I didn’t expect to have SO MANY people from both sides of the aisle agree with my last post!!! It may not seem like much, but I think that’s a sign of actual progress as a community! That is honestly amazing to me!
Awesome, thanks for sharing your work with everybody.
Regarding the seed modifier, this looks great and it appears to serve the same function as the ‘generate/randomize match’ button I was asking about. I have entered a few different values and see a new MMR simulations being generated. Player MMR and SR data are not changing though, and neither are the direct input values. Is it possible to generate new numbers in these parameters when the seed is changing?
Ah, changing the random seed is only to ensure that the simulation itself is consistent. Changing the seed changes the result.
Also keep in mind that the direct inputs are only there to build/account for one team. If you want random teammates, just remove the values and random teammates will replace them!
Do note that this will also change the calculation.
This might be slightly off topic from me, but I think we can all agree to give credit where it’s due, you did a fantastic job on this. Just because the sides disagree, doesn’t mean we don’t see both sides, we just look at it differently
What a nice thing to say! Certainly not off-topic.
I agree, Nano’s tool is really cool and I think it will help players understand the concepts of probability and mathematical statistics that we are discussing.
Believe it or not, I do try to look at both sides objectively, as I said before, cudos to you for the work you’ve done (and that video was fantastic).
I know I come off as biased, but from my point of view, the matchmaker does work as intended (or the best it can, alts, smurfs, throwers, new players counted). I would more than willingly try out an SR only based system, but by that same token, I would also like to try an MMR only based system, so we have a metric to base the tests on.
You seem like a very curious individual, and surely you would like to weigh the pros and cons of each system.
my concern is what if you’re MMR is above average that you’re getting bad team mates and this proves even more detrimental than being normal MMR to better team mates.
reliable team mates is better.
in a way, its like being punished for being good, or being punished to be able to play many heroes
Your entire post can basically be dismissed as speculation, since you purport to know more than you possibly could and you have no means to verify or confirm what you’re saying. Yet you and so many others want to make truth claims, when you don’t (and can’t) know the objectives and goals of the matchmaker, what it’s optimized for, how the algorithm works, how or if Blizzard values fairness or competitive integrity over, say, profit/engagement statistics, whether or not key premises on which your argument is based are factual, whether previous developer statements still apply, etc. etc.
It seems no one is comfortable acknowledging that they know very little about a closed proprietary system in favor of rampant dabbling in the availability bias fallacy: which gives undue attention and importance to information that is immediately available (piecemeal conclusions, first hand observations, unquestioned assumptions, anecdotal tidbits from others, ad hoc/post hoc theory-crafting, etc) while minimizing or ignoring the importance of key evidence that clearly exists but is not easily accessed – in this case, the actual runtime code for Overwatch itself.
You’re free to dismiss it. I don’t mind at all. Your post’s content is akin to taking the stance of solipsism in philosophy, though- which is to say that you seem to think that we can’t meaningfully discuss that which is unknowable. The complaint you throw against my post could be applied to both sides of the row for matchmaking issues. It is a fair criticism, no doubt, but it doesn’t really further the conversation.
I said it long ago, and I’m happy to reiterate it for you, I think we have absolutely no way of knowing what is actually running under the hood. They could be gaming the sweet bejeebus out of human psychology in a ridiculously cruel manner to achieve arbitrary engagement metrics and we wouldn’t have the data needed to prove that’s what was happening. Conversely, it could be an entirely unbiased system that just has some flaws baked into it that are all completely unintentional from a design perspective.
That disclaimer aside: I simply don’t think it’s productive to operate under the assumption that disregarding the information provided to us by the developers is the right call. The moment you accept that the developer posts we’ve been provided with over the years are either irrelevant due to a change we weren’t informed of or dev posts were written in a way that was intentionally misleading, every hypothesis presented becomes nothing more than speculation and conjecture.
Many of my posts can be summarized as follows: “I assume what we’ve been told is true and still applies to the current game despite a long window with no additional information being provided. If true, then X would happen in scenario with variables A, B, and C.” The moment you reject the initial presupposition, then, of course, everything that follows is wasted breath.
I hope that clears it up at least a bit. I’m not claiming to know truth or to have insider information. I have no issues or discomfort associated with saying that not only do we not know how it all works, but that without additional information that we currently do not have access to, we cannot know how the matchmaking system works in its totality.
I’m interested in recent comments about solipsism and known unknowns. This is why I am so focused on the developer statement by Scott Mercer from 2016, where he talks about MMR, forced 50% match odds, and discriminating between players based on their performance statistics. Because it is the only real statement of substance that we have about Match Making Rating from a primary source, the Lead Designer of Overwatch. It has been scrubbed from the internet with the rest of the old Battlenet forums…the quotation resides in this thread alone.
Though I have always felt that such a telling statement should be enough, I had to look closer at Activision/Blizzard’s patents related to matchmaking, when I started to receive tips from contributors to this thread. I understand that line of inquiry has led me into what some would consider conjecture…but I prefer to think of it probabilistically, in much the same way that Blizzard profiles their players with MMR.
What are the odds that Overwatch’s matchmaker is completely different from and uninformed by Activision/Blizzard’s 2015 Matchmaker patent? I would suggest the number is somewhere between 0% and 1%, based on the information that is currently available to us which is:
A corroborating statement from Overwatch’s Lead Developer.
Patent filings from the release year of Overwatch, containing descriptions of invention that align with developer statements about Overwatch’s Matchmaker, and the SR/MMR system.
Activision/Blizzard’s history of inventions related to matchmaking, each of whose reference lists are a mile long, referring to patents like Microsoft’s 2006 “Bayesian Skill Scoring Framework” invention, which is strongly analogous to Activision/Blizzard’s matchmaker if not an actual predecessor.
The long corporate history of unethical and ruthlessly addictive game design by Activision/Blizzard, as seen in titles like World of Warcraft.
The class action lawsuit for workplace misconduct and mismanagement, and the recent suicide of that poor woman who was abused in Activision/Blizzard’s offices. General signs of institutional depravity and disregard for ethics.
I could go on, and I will in future edits…but these are the things that come to mind right now. At a certain point, I think we have to respect our own state of knowledge and stop doubting the facts because of Activision/Blizzard’s mystifying silence. Their failure to reply to a thread that has run for 4+ years, receiving almost a hundred thousand player comments and views, is just that: a failure, and not a contradiction to my argument.
It likely isn’t completely different, but just having a patent and putting every aspect of said patent to use aren’t one in the same. For example, they had a patent that cosmetics would be used for matchmaking purposes, but went on record to state they didn’t use that in practice, IIRC.
That aside, there is still going to be disagreements around matchmaking at a fundamental level between folks like you and I. I actively prefer the idea of trying to balance matches with the available data (even if only SR ends up being used), and you seem not to want that. Or have I got you all wrong on that part?
Edit:
Kaawumba still has threads on this forum that quote the old posts, FYI. Not as good as the original colored text, but unless you think he’s doctored the statements instead of posting them verbatim, it could be useful for you. I can find a thread with some of them if you’d like.
Edit 2:
Solipsism is one of those really cool things that someone occasionally latches on to way too hard for their own good. We can’t know for certain we aren’t a brain in a vat or inside the matrix (or whatever other creative example you want). For instance, I can’t know for sure you are a real person that isn’t just a figment of my imagination. However, it wouldn’t be productive for me to assume that because I cannot know whether or not you exist, that it follows I shouldn’t invest time/energy/thought into our engagements.
I believe the term the “hard solipsism” is thrown on the people that go completely nihilistic with the stance, for reference.
You’re right. Activision/Blizzard’s 2015 Matchmaker patent includes many troubling details of invention, including plans to discriminate against players based on gender, income, and residential location. But there is no telling which plans have been implemented, in any given Activision/Blizzard title, because the corporation is keeping players’ terms of use a secret.
However I am not taking issue with mere details of Activision/Blizzard’s Matchmaker invention; I take issue with the invention itself and its stated purpose. I think that discriminating against players and segregating them from each other based on their relative skill is unethical. At the very least, I think it is unethical to do so while purporting to rank players based on their skill. Players’ chance to advance in rank depends mainly on their ability to win games. Match Making Rating profoundly reduces skilled players’ odds of winning a given match.
I agree that only SR should be used for the purpose of matchmaking. However I do not think it should be used to balance/handicap matches, and in fact it is not SR but mainly MMR which serves this purpose. I think that Competitive Play should be solo-queue only, which would allow fast and fair matchmaking.
I believe it was Kaawumba who originally pointed me to Scott Mercer’s statement on the Battlenet forum, which is the basis for my argument. But he didn’t make anything of Mercer’s statement himself, rather suggesting that it proved the SR/MMR system to be innocuous. I think it proves exactly the opposite. I suppose people can see the same thing very differently.
I believe this best summarizes our disagreement. Regardless of the metrics used, I would want matches to be balanced. I want this to be a thing in hopes that it would increase the quality of individual matches. You do not seem to hold that view.
People absolutely can interpret things differently. Glass half full vs half empty is the most digestible example I can provide.
Here is the thread I mentioned where Kaawumba saved some blue posts on the off-chance you might find it useful.:
As a final, unrelated note, I may be going off the grid for a long while again. I find myself playing Overwatch to “scratch an itch” rather than actually enjoying the game. I think I’ll be uninstalling it soon. From there, I’ll probably abandon forums, too.
If I disappear, I wish you all the best. Keep the open conversations going. It’s been honor to contribute to the back and forth, even though what I’ve provided is modest when compared to most of you. Take care, you beautiful people.
Yeah, that is the major point of difference between our positions. However, I would say that impartial/fair matchmaking is the best way to ensure match quality. MMR is designed to find out the difference in players’ skill, relative to others at their SR level. With that state of knowledge, the matchmaker balances/handicaps teams to have equal chances of winning. But that comes at the expense of the most skilled players in every match, reducing their chances of victory and advancement of rank. I don’t believe the ladder functions under these conditions.
I believe the dysfunction of the ranking system is caused by Match Making Rating, and that players can sense the wide disparity of skill between players in every rank. This results in cognitive dissonance, which causes toxic player behavior and gives everyone a general sense of unfairness that is unidentifiable, because of steps that Activision/Blizzard has taken to keep our terms of use secret from us.
I understand the need to do this. I stopped playing Overwatch completely (again) after producing my video. I attend the forums only to participate in discussion about Match Making Rating, as a matter of truth and justice. I hope you’ll subscribe to my YouTube channel which has new episodes coming out soon, moving on from videogames to other subjects.
I wish you all the best, and thank you for participating. You have elevated our discussion with your thoughtful and well-considered comments! Fond regards, Sheevah.
