Algorithmic Handicapping (MMR) is Wrong for Overwatch

It has been proven in several threads.
Refusal to self-inform = personal denial. You’re not fact checking, you’re just blindly denying. That is offtopic and inapropriate for this thread.

Gambling goes against the BBB and rigging goes against esports. So they are exposing themselves to a class-action for misrep and damages. Not to mention a lot of bad press when the case forces an inspection of the algs :stuck_out_tongue:

A dev statement on record, an active patent which mentions OW, statements from the original authors on file, a screenshot that shows the game.

That’s enough for most reasonable people, so we’ll go ahead with the discussion. You are in active denial if you don’t adjust your belief function. No amount of evidence will sway your blind devotion to the corporation.

Neither could they. Case in point: the layoffs, 0 ladder maintenance, fairplay integrity under the bus, etc. It’s entertainment, they favour rigging over engagement and sales.

Stop damage controlling for the company. It’s offensive to customers. The community deserves consumer protection and awareness. “I don’t believe you” isn’t enough anymore - you’ll have to show evidence that the patents aren’t in use, that devs have said they don’t mmr for 50/50, etc.

Pro tip: Stop denying, start fact-checking.

I’d like to take a moment to thank the actual contributors to this thread for their hard-work in exposing the issues. Whether or not you believe them, whether or not 100% true. It’s highly sus, and it’s a great technical debate about the workings of rigging and why/not to ever have it in esports.

1 Like

I really don’t like a competitive ladder that only takes winrate into account.

You either win or lose, those are simple rules… games have grown more complex, though for a good reason of course.

2 Likes

Then tell us what we’re being scored, rigged, and paid out on?
I really don’t like a team that backs a hidden rigged competition mode and labels it an esport. No place for that kind of misrepresentation.

1 Like

Well, I love it when a system is made to evaluate my performance and puts me in a matchmaker that knows what I’m capable of.

This way I’m put in a fair match because the performance is tracked of every player in the lobby, this way everyone has equal performance… at least somewhat close.

I think this is way more stable than winrate, there are many factors that can result in players being boosted unintentionally and this way the ladder keeps its integrity.

2 Likes

Incorrect. Because forcing the matchmaking unforces the ranks. You can’t have it both ways. Either you get proper labels or you get fake rigged “close” matches (which aren’t actually fair).

In real competition and real ladders your rank is all the assurance you need for “fairness”. It’s a hands-off approach, which means you can stomp, be stomped, streak, etc. But if SR labels/classifications are remotely fit-for-purpose, than that’s all you ever need to ship a match AND have fair rewards.

Don’t you want to know what is being tracked? Don’t you value your data trail and question what the performance criteria might actually be?

As long as this is not done through rigging and handicapping - and is publically auditable (every sport needs to know the rules and evaluation criteria). Otherwise the ladder is of flawed design, or intentionally an anti-competitive scam.

1 Like

Oh but they are fair, my experience indicates so.

I think it’s still fair though, all they do is track correct performance and i’m happy with it.

People who don’t agree with this, are in my eyes just blaming something so they don’t have to take the blame.

Not really, i’m not playing to satisfy the exact things it’s tracking… i’m trying to play a game here. Though if it were only based on winrate then it wouldn’t be much fair to me. This would also heavily affect the MM.

I think these threads give me enough information to work with: How Competitive Matchmaking and Rating Works (Season 27)

1 Like

I have updated the information on my polls, and appended my original post with analysis of Activision’s 2015 patent for a matchmaking system. I hope that fans and critics of the original post will re-read it, to familiarize themselves with relevant details of the patent. Activision’s Matchmaker description aligns with what we have heard from Scott Mercer and other Blizzard reps about the function of Overwatch’s Matchmaker.

Overwatch players, and gamers in general, deserve to know their terms of use in ranked competitive play. Activision’s Matchmaker patent reveals those terms to include skill measurement for handicapping. Concerningly, it also includes forms of user discrimination based on demographic information, specifically including and not limited to gender, income, and residence.

Yeah that’s not going to cut it. We have objective, well-defined measures of fairness that abstract away from people’s subjective interpretation of “fair”. You can’t patent naive, unadulterated fair. You can patent rigged mechanisms like slot machines and all the sleeze in the casino/gambling business.

The system doesn’t let you properly compete. It doesn’t tell you how you’re scored or show you that score. It rigs. They’re blaming a rigged design that takes away their rightful ownership of their data and results, a rigged sham that takes away their control and influence on a match.

It’s puke-worthy. And I hope they have to close it down.

1 Like

If what were only based on win rate?

Because then it isn’t tracking any of my performance, winrate is decided fully by a loss or not.

Factors like players on your team can affect your win, hence why SBMM is useful for lower rankings that are still improving individual wise instead of fully team based.

Possible is that ranks will even be further split apart.

1 Like

I mean… maybe if this was OWL the devs would care… but it isnt.

1 Like

It doesn’t matter if they care or not, the current system is giving me matches based on my performance.

I find this fair, and it helps against having all players from Bronze-Platinum dropping to Bronze-Silver whilst the Diamond+'s will stay exactly there creating some empty ranks in the middle.

2 Likes

You can’t have two seperate ways to count in the same ranking system. That in and of itself makes it questionable integrity. I don’t care if it’s performance based or win based of your ranking. It just needs to be one. The cream will always rise to the top.

2 Likes

Winrate is technically all you need to ship a match since it’s basically what the ladder already does (no-reset per-account winrate).

SBMM is garbage. I think you’re referring to PBSR, which is all you would ever need. You can have a system that ships 12 random people by SR band, plays out a match, and rewards your standard +/- SR. You are then judged on your ability to convert wins for/against others who are supposedly at the same ability. You can add in some PBSR in the lower ranks to help kick people out if they truly don’t belong.

But it isn’t labelling you properly. Which is entirely unfair. You’re equating the “closeness” of the match as your performance reward, when it ought to be an accurate and precise label (payout by SR, chevrons, and pixels).

This is on the right track. If SR is fit-for-purpose in terms of scoring and ranking and paying people out - why is it not used for matchmaking? Because they don’t know how to build ladders at this company. They build amazing worlds, do great artwork, and make things engaging and replayable. But competitive integrity just wasn’t hired on.

1 Like

Even so i find it more balanced.

System is fine as is and anyone who blames it is just bad :ok_hand:

1 Like

What you find doesn’t matter. You’re really not important compared to a large population backdrop.

The only objectivity we have atm is examining the math behind their model. And the math says it is NOT proper and CANT be working. You can 100% blame the system if it is using math that errodes rank integrity and is anti-competitive.

2 Likes

I never say I’m not where I belong in my ranks, so if you were trying to take a parting shot… you didn’t. I’m saying you essentially have two sets of rules, in one competitive arena, which for integrity of the game… would be like saying the NCAA Mens, and NBA are competing for the same title with their current rules.

They play the same game, but it isn’t the same rules and regs. so… you either make it all the same rules so we can all play the same game, or seperate them completely.

2 Likes

You aren’t either, lets make that clear to start with, I’m not important either.

Once again, you DO NOT have the math needed for this, you can claim you do all day long, but the bottom line is, you don’t, you don’t have the inside works for this game, stop lying.

The only reason you blame the system is because it’s not working the way “you” want it too, other players, I would say the majority, have no issue with the current system, these forums do not paint a true picture, since 99% of the active players do not use them. Sure you might get 100 players here on the forums that agree, but there is still tens of thousands of players that do not share this idea.

2 Likes

This makes it different games, not the same, once you change the rules and regulations you change the entire aspect of the game, they might be similar, but not the same.

No, we have one set, SR is a visual representation on your MMR, SR’s basically a meaningless number that players flaunt because theirs is higher.

1 Like

It’s amazing what flat earthers will go through to deny the science.

Less gaming, more math. Knowledge is power. We know it’s trash, because it has to be, from first principles and public facts. That’s without leveraging the existence of patents or dev statements btw.

Things like time-since-last-reset, alt disruption rates per-alt, normal pdf error rates, entropy and signal/noise growth rates for various N levels, etc. They are all useful information for the models. These things let us algebraically bound what the ladder can and can’t be. You don’t need detailed data or inner workings for this.

It’s rather basic, and comes up all the time in systems identification topics. An undergrad is already qualified enough to talk shop on these issues.

1 Like