The math behind DbD's MMR is just a single "if" statement (900h vs 11,000h lobby analysis

That is exactly the problem I’m pointing out. If a player has 7,000 hours of actual mechanical experience, but the internal system only "sees" them as having 1,000 hours, then the system is fundamentally flawed and data-blind.

You can't ignore 6,000 hours of muscle memory, map knowledge, and mechanical mastery just because an arbitrary "score" hit a ceiling. From a programming perspective, if your input data (actual skill/experience) is vastly different from your processed variable (the hidden score), your output (the match quality) will always be garbage.

The fact that the system treats a 10k-hour veteran and a 1k-hour intermediate player as a "perfect match" isn't a feature—it’s a massive failure in data granularity. It confirms that the "hidden score" doesn't reflect reality once you pass a very low threshold. We shouldn't be making excuses for a system that chooses to be ignorant of the most important metric in a competitive environment: actual time spent mastering the mechanics. The core issue is that the algorithm relies on almost binary kill/escape data. It completely ignores what actually happens during the match, which is just bad design. Treating a 10k-hour veteran and a 1k-hour intermediate as a "perfect match" isn't a feature—it’s a massive failure in data granularity. It confirms that the hidden score doesn't reflect reality once you pass a very low threshold. We shouldn't make excuses for a system that chooses to ignore actual time spent mastering the mechanics.

You make a fair point in theory: a system should use highly detailed skill metrics instead of raw playtime. But let's look at the reality of how the skill ceiling in this game actually scales.

Saying that a 1,000-hour player is on the exact same playing field as an 11,000-hour player just because "they both know the game well enough" is like saying a Junior Developer and a Senior Systems Architect should be given the exact same tasks because they both know the basic syntax of Python. (I'm learning programming, which is why this comparison came to mind.)

At 1,000 hours, you absolutely know the mechanics, the perks, and the general map layouts. But at 11,000 hours, you have ingrained muscle memory for every single tile permutation, edge-case mind games, and pixel-perfect hitbox manipulation. You simply cannot bridge a 10,000-hour gap of pure, mechanical experience just by "knowing how to play."

However, you hit the nail on the head with your last sentence: "The game should have collected data about your skill level." The tragic part is that it doesn't. Because the matchmaker's data collection is strictly binary (kill or escape), it completely fails to recognize the massive mechanical gap between a 1k-hour intermediate and a 10k-hour veteran. Since the internal variables are useless and blind to actual skill, playtime hours remain the only raw data we have left to expose how broken the matchmaking actually is.

You are asking the exact right questions. How do we actually fix this architecture instead of just pointing out the bugs?

Regarding your idea about a seasonal MMR reset—you make a solid point, it could actually help flush out stagnant data. You mentioned the fear of smurfing, but let's be honest: smurfing is already a massive issue in DbD anyway. A structured seasonal reset might actually be much healthier than the permanent, broken data buckets we have now.

Another practical fix would be a return to the old visible Rank system, but completely redesigned with strict enforcement and hard caps. The old ranks failed because they just measured monthly playtime. If BHVR built a highly accurate, heavily monitored ladder with strong thresholds—meaning once you reach the "veteran" bracket, you are hard-locked and literally cannot be matched with casuals or beginners, no matter how long the queue takes—it would solve the mismatching instantly.

And that brings us to the hard truth about queue times: in every serious competitive game (CS, Valorant), the top 0.1% of players always have 10+ minute queue times. That is the price of being at the top. BHVR is just terrified of 10k-hour streamers complaining about queues, so their algorithm intentionally sacrifices average players as "server backfill" to give those squads fast matches.

The fix isn't impossible. It just requires granular data tracking (chase times, not just binary kill/escape) and strictly enforced, hard-capped brackets.

You bring up a very valid concern about code complexity, but from a software architecture perspective, it's actually not impossible to manage at all. In fact, the framework already exists in the game.

You don't need a million if statements. Modern game engines use event listeners. Dead by Daylight already tracks almost everything you mentioned. When you bodyblock for a teammate, the server registers a "Protection Hit" event. When you run the Killer around, it tracks "Chase Duration". The server already knows exactly how much of a generator you repaired. The coding failure isn't in gathering the data; it's that BHVR actively decided not to plug those existing variables into the algorithm.

Regarding your point about "bully" tactics—handling this mathematically is incredibly simple. We know exactly how a bully squad operates: they focus entirely on harassment (chases, unhooks, bodyblocks) and ignore the main objective (generators). Since the server tracks all of this, the algorithm would simply see a massive spike in 'Altruism/Boldness' metrics and a flat zero in 'Objective'. If a proper MMR formula requires a balanced score across multiple categories to rank up accurately, a squad that refuses to do gens wouldn't break the system at all. They would just be scored exactly based on their specific variable inputs. There is no complex philosophy or impossible coding here—just basic data weighting.

Instead of building a weighted mathematical formula, BHVR threw all that rich, granular data in the trash. They opted for the laziest possible boolean logic: if (escaped) { MMR++; } else { MMR--; }. It's not that a detailed system is impossible to code. It's just that relying on a binary outcome requires absolutely zero balancing effort from the developers.

You completely misunderstood my position. I never claimed to have high MMR. In fact, I play pretty averagely, and based on my mechanical skill, my MMR should be low. If any experienced 11k-hour veteran actually watched my gameplay, they would be shocked that the system paired us together.

And that is exactly the fatal error of this algorithm. It completely ignores what actually happens during the match—the very data that would reveal my true, average skill level—and relies solely on a 1/0 binary outcome. Because of this blind spot, it artificially inflates my score after a few lucky escapes and randomly throws me against players at the absolute highest MMR cap.

In any other serious competitive multiplayer game, this would never happen. In Valorant, a low-ranked player doesn't get matched against a Radiant just because they had 2 lucky games and their hidden MMR temporarily spiked. The system prevents that type of mismatch.

But in DbD, as you literally just admitted, once you cross a low threshold, "there's no upper limit to who you will face". That isn't a functioning matchmaking system; that is a single, massive data bucket.

We aren't asking for easy wins or to be "sacrificial lambs". We are asking for an algorithm that actually measures mechanical skill instead of using average players as server backfill for top-tier squads.

Thank you for bringing this up. This piece of history is the absolute smoking gun that proves my entire point.

From a software development perspective, this is the worst kind of technical regression. It proves that the current broken matchmaking isn't just a result of incompetence; it was a deliberate downgrade. BHVR actually had a strict, functioning algorithm, but because top-tier content creators couldn't handle playing against their own skill level (which made it harder to farm easy clips for YouTube), the developers intentionally nerfed the architecture.

They literally rolled back a working patch to appease a vocal minority of influencers.

And what is the result? We are left with this "no MMR" system where statistically average, 900-hour players are deliberately fed to 11,000-hour veteran squads. BHVR is actively using the casual and average player base as disposable server backfill, just so Twitch streamers don't have to wait in queues and can have "chill" games stomping weaker opponents.

It is an absolute joke that a competitive multiplayer game balances its core matchmaking variables around the comfort of content creators instead of mathematical fairness.
I know about these changes, but that doesn't change the fact that the current system doesn't work. That system was also poorly designed, which is why there were complaints. But if they had done it right (it's possible, you just have to want to), I think both sides would be happy. Weak killers vs. weak survivors. Good killers vs. good survivors.

I understand the argument about someone spending 8,000 hours throwing pebbles, but in data analytics, you don't design a core system around extreme statistical anomalies. You build it around the median, and median playtime strongly correlates with muscle memory and game sense.

But let's talk about your point regarding the 2v8 mode and the absence of MMR. You feel it's a "better system" because it replaces a rigged, frustrating algorithm with pure, unadulterated chaos. However, from a balance perspective, 2v8 without MMR isn't a solution—it's just a casino roulette. Let's break down the actual scenarios:

2 weak Killers vs 8 weak Survivors: The result is pure RNG. It could end with 0 escapes, 4 escapes, or all 8 escaping. There is no predictive logic or balance here.

2 weak Killers vs 8 good Survivors: The outcome relies entirely on Survivor hubris. If the good survivors make careless mistakes, nobody escapes. If they play optimally with zero mistakes, the vast majority escape easily.

2 good Killers vs 8 good Survivors: This is the ultimate roulette. The match is dictated entirely by which TIER of Killers are picked and the exact micro-mistakes made by either side during the chaos.. Here, the outcome depends on what kind of killers they are (which TIER killer) and on the mistakes made by the killers and survivors.

The reason 2v8 feels "refreshing" isn't because DbD is incapable of accurate matchmaking. It's simply because playing a chaotic party mode with zero matchmaking is temporarily less frustrating than playing 1v4 with a broken matchmaking script that uses average players as backfill.

We shouldn't accept that the game will "always be a crapshoot" just because BHVR refuses to code a granular, properly weighted MMR system. DBD is suitable for matching players appropriately; no one is talking about turbo precision, only about matching players correctly. This is necessary for game balance. Otherwise, nerfs and buffs on both sides are pointless, because what's the point of a survivor getting buffs if they're weak and encounter a skilled killer who's been nerfed? If the survivor wins, it's not because of their skills, but because of those nerfs and buffs. If they lose and the killer wins, the killer will win only thanks to those nerfs and buffs, not thanks to their skills. And then they'll write “noob, uninstall this game.” That's not what multiplayer games are about.

That is a fun analogy, but you are completely missing the scale of the math here.

You are right—playing baseball since you were 6 years old doesn't automatically make you an MLB professional. But if you take a 30-year-old who has played amateur baseball every weekend for 24 years, and put him in a match against someone who just learned how to hold a bat a few months ago, the veteran is going to absolutely obliterate the newcomer.

No one is claiming that 11,000 hours guarantees you are a professional. What we are saying is that a 10,000-hour gap represents a massive, mathematically undeniable advantage in muscle memory, map knowledge, and mechanical efficiency.

If a matchmaking algorithm looks at the 24-year veteran and the 6-month beginner and says, "These two are a perfect match because they both won their last game," the algorithm is fundamentally broken. It is confusing a basic binary outcome (a win/loss) with actual mechanical skill.

Since BHVR's hidden MMR doesn't track granular skill, playtime remains the best heuristic we have to expose these absurd mismatches.

Which doesn't work anyway… Because you'll often encounter, for example, 10 or even more unfavorable matches.

You are creating a false dichotomy. You assume the system must either be a simple binary 1/0 (escaped/died) or a one-dimensional "chase timer." In modern software architecture, that is not how skill-based matchmaking works.

A professional MMR system uses weighted multi-variable analysis.

If a Survivor runs the Killer for 3 minutes (chase metric), completes 2 generators (objective metric), and saves 2 teammates (altruism metric) but dies at the end, a competent algorithm doesn't just say "0/1 - Fail". It weighs those high-value inputs against the outcome. Currently, the system rewards a player who hides in a locker for 10 minutes and sneaks out the hatch, while punishing the player who actually did all the work. That is a bad incentive, and it's exactly what we have now.

As for your "explloiting the metric" argument:

The "Pallet Pre-dropping" point: A weighted system can easily account for this. If a survivor pre-drops 10 pallets in a 60-second chase, the "resource efficiency" variable would be low, balancing their MMR gain.

Killer Design (Nurse/Blight): This is exactly why we need granular data. The system should know that a 30-second chase against a P100 Nurse is statistically more valuable than a 60-second chase against a Trapper.

The objective of the game is indeed to escape, but the skill required to reach that objective is what the MMR is supposed to measure. Right now, the system only measures the result, not the skill. If you think a system that can't tell the difference between a mechanical master who sacrificed himself and a beginner who hid the whole match is "working," then you are settling for a massive failure in data granularity.
Your logic is self-contradictory. You admit that a player can "run the killer for four generators" and still claim that measuring that skill would be "unreliable".

From a technical standpoint, if one player is occupying the Killer for the time it takes to complete 80% of the primary objective, that player is the primary driver of the win condition. In any balanced system, that should be rewarded.

The "bad incentives" argument also fails for two reasons:

Reality check: Survivors are already pre-dropping pallets constantly in the current meta. Acting like a better MMR system would suddenly "break" the game with pre-drops ignores the fact that it's already a standard low-skill tactic.

Weighted Math: A well-designed system would simply treat "Time in Chase" as one of many variables. It wouldn't mean you "farm" MMR just by running around; it means that if you perform the hardest mechanical task in the game (looping) which allows your team to finish the main objective (generators), the system acknowledges your skill.

Currently, the binary 1/0 system treats a mechanical master who carried the team but died at the end exactly the same as a beginner who did nothing. That is the real failure in data granularity. If you are good at looping, your MMR should reflect that skill slightly—not as a "free win," but as a recognition of your mechanical input into the match outcome.

It’s not about "farming long chases"; it's about the system finally opening its eyes to what is actually happening during the trial instead of just looking at the exit gate.