What if Blizzard would remove MMR and rating was simply +25 or -25 if two equally rated teams play against and if the enemy team is higher/lower rated rating grows or drops more.
What I find wrong is that two teams with equal MMR can just go for 1 win - 1 lose and still gain rating together. Make it simple. Wintrading won’t be viable because people will have to sacrifice their ratings in order to give someone else free rating. And then it might be harder to climb back.
I know that MMR was invented so high rated players won’t play against newbies but it happens already on alts. What’s the point. Dota 2 has similar system but way simplier - win is +25, lose is -25. The only difference is that seasons are longer and ratings are higher - top rating currently is 11 k while mediocre player start at 2k.
MMR in WoW caused more problems and created gray area for exploits than benefits it gives. Maybe it’s time to try something else?
I’m not convinced that MMR is 100% the problem but I was thinking about it recently. What is your idea about it?
To begin with, Blizzard haven’t “invented” anything about the rating system, per se. What they’ve done is modified an existing rating system. And no, this isn’t the Elo rating system (per se). The only time Blizzard used the Elo rating system in WoW was back in Burning Crusade, but with a very tiny modification.
The current rating system in WoW is based on rating deviation, which the Glicko RD system introduced. (Which is kinda based on Elo, but it isn’t the Elo system.)
The purpose of rating systems as a whole though, is just to quantify your hypothetical “skill”. Every person, even the ones who hasn’t played the game yet, have a corresponding numerical value in a rating system that represents their ability to play the game. The mathematical side to rating systems, is how to get players to reach their hypothetical “true rating” in the best way possible.
Now, it’s theorized that it’s impossible to ever truly reach that rating, so the mathematical proofs of rating systems are judged primarily based on 2 factors to determine their legitimacy:
How close to that hypothetical “true rating” a player can reach in the system. Which in theory means the limit of the accuracy of the rating system as a whole, for how close a player can ever get to their “true skill” value.
In other words, it refers to the way it’s designed mathematically, to go from a player who knows nothing about the skill value yet, to get as close as possible to that numerical value, and how that impacts the representational value of the number, via the use of the way the rating system is designed.
So the other way a rating system is judged, is how fast a player can reach that “practical limit” of the rating system’s accuracy. As in how many games it’d require in the mathematical proof.
While Blizzard spinned it in their announcement of the separated MMR from the CR while preparing for wotlk that it’s for the sake of making the better players play on ratings “better suited for them”, that’s just their way of dumbing it down for the masses.
To matchmake players based on rating, which WoW has always done, is simply a means to an end in a rating system, and how it’s mathematically designed to get players to their “true rating”.
The thing about why it’s theoretically impossible to ever reach that true rating that perfectly represents a person’s skill in something, is because people always change. Your ability to play at one point during a day can be different from your ability to play at a different hour during that same day. Our brains, energy levels, ability to focus, and basically everything no matter how small you might think it is, affects the way we make decisions, the way we react and as a result also the way we play.
This is what repeated practice is partly for, it helps stabilize our ability to do something. It’s part of why athletes do it, it’s part of why chess players do it, and as a practical example you can see the same effects in WoW. A veteran PvPer is much more consistent and more versatile in their ability to play, compared to a beginner, even if they momentarily might end up making the same decisions sometimes.
But all in all, your hypothetical true skill will vary from morning 'til night, and ofc from day to day. It’s never truly static.
Fun fact:
It’s easier to control ladder inflation in WoW’s current rating system, than it is in the Elo rating system. Because the areas in the system where “new” rating points are created, is in many more places compared to the Elo rating system. In the Elo rating system, new rating is only ever created at the time a new unit joins the ladder. In WoW’s current system, you’ve got new rating points being created all over the place, with people winning more than their opponents lose.
It can deflate by doing the opposite, to lose more than the opponents win, but that doesn’t really happen as often. So by simply manipulating the matchmaking metrics, and they can even isolate specific rating ranges to manipulate the matchmaking metrics, Blizzard can control the ladder inflation as they see fit. With nobody noticing a thing.
In the Elo rating system, since you’d win what the opponents lost all the time, the actual points people would fight over was only ever created with each new unit injection into the ladder.
People back then confused this for inflating the ladder when people would ditch their old team to create a new team over and over again, but that isn’t exactly true.
The team was a single unit injection into the ladder, which is why it had to start on 1500 when the CR was the MMR or else everyone would be stuck on 0, but if players would delete a team to create a new team, then they’d still only be representing the same amount of rating points being fought over. It’d only inflate if people would be jumping between different teams, basically making the same characters represent more than 1 unit injection.
So for a team to delete their previous team and create a new one, think of it more as resetting their placement in the ladder, instead of it actually inflating the ladder, back then during TBC.
The benefit is on the design side. They can easily control the ladder inflation when they choose to, and thanks to the massive matchmaking pool as well as how what used to be 1 unit injection of 3 characters (in 3s) became 3 unit injections, they can get more players to reach their desired rewards.
Some people will always be excluded though. That’s just the way it is.
TL;DR:
It’s about the nitty gritty of what rating systems are for, and how that correlates to the OP’s post.
Simply put, the premise the OP stated is a misunderstanding, which is partly why the conclusion is wrong. Good job to anyone who actually manages to read through the entire wall of text though.
PS:
Wintrading wouldn’t be a problem if Blizzard would actively police their own ladders. Not everything needs an automatic system to solve it, nor can everything be solved with automatic systems. Sometimes manual labor is required to achieve certain results, such as effectively enforcing their own rules. Because if they do it well enough, it gets a ripple effect where it creates a reason to fear punishment if anyone should even try it, thus the mere perception of the likelihood to get punished can work well enough as a deterrent. Which is what they’ve brutally failed with for so many years now, and gradually gotten worse at.
As for the issue of new rating points being created by wintrading so they can all get higher rating to secure r1 by doing so, that’s related to the core of the ladder inflation controls.
Another way to put it, is that they are probably already aware of this systemic weakness, but to change the core would affect the ladder inflation as a whole. Which is why actively enforcing their own rules would let the ladder function better, without needing to change a thing.