You’re going to need a lot more data and a very controlled experiment to prove this.
Anecdotal, Sample Size, and Confirmation Bias are all expressions that come to mind here. What you are presenting is not evidence. It’s what you feel is happening, but actually has no basis in fact. The data just doesn’t support your conclusion.
This is your argument? Ad hominems? How about presenting some evidence? Show us the data mined code for this algorithm. Show us internal communications within Blizzard that prove this has been implemented. Show us detailed logs of controlled experimentation with collected data from a sample size of 10,000.
The parameters, as stated by the developers, is simply the following:
MMR while you have the star multiplier;
Rank when you no longer have the multiplier;
MMR at Legend.
No you didn’t. You just described a small part of a proposed experiment. Show me the controls you put in place: what rank, what deck, what time, what day. Show me the 10,000 data points for each deck.
This morning I played 2 matches with my Druid and faced a DH and a Mage. I played one match with Weapon Rogue and faced a Druid. I played one match with Control Warrior and faced a Shaman. Is that enough data to disprove your theory?
Within your data collection what are you doing to minimize out side variables for the collection?
I don’t have any studies, but I have observed that there are changes in the meta around half hour to hour increments which would be inline with others players time zones. And I have also found that about 2 hours after I eat dinner the meta is pretty often not so punishing for control. What does you data collection do to count what other decks are being played at that time?
Just because on the vs reports shows an overall meta it does not mean during the entirety of the time data was collected the meta was 10% this, 12% that, etc. There is no way to guarantee you face any certainty deck, but there is a way to guarantee you face conter decks more often than not. That is by making your deck to only conter one deck.
Sure, I have seen times were I change decks and still face bad match ups, but I play my poorly designed homebrews so I really can’t expect much and I can’t say with any certainty matchmaking was not random.
You can do it yourself so you cant deny the prove isnt real, go do it and you will see the same results
You can play druid and face 30 druids in 40 games and then play fatigue warlock for example and suddently its max 2-3 druids in 40 games, its obviously rigged you gotta be blind to not see it
I know it’s just an example you happened to give, but according to hsreplay, quest handlock has more recorded games against Druid than any other class. Well, since the last time their data reset (anyone know if they reset it every patch?).
You can use this info btw. The numbers won’t match exactly of course do to deck popularity, but the percentage of games should be pretty damn close.
Pick a deck. I believe you use Miracle Priest as an example so I will too. You think it gets unfairly matched against Mages with devolving Missiles (Ping Mage is the only list that runs Missiles that I know of). Quest Warrior has an extremely similar spread against Mage decks. So does Ramp Druid. If Miracle Priest decks were being unfairly matched against Ping Mage more than other decks, it would show.
Edit - I also don’t see the point of rigging EVERYTHING. Why would they have to rig MM, to queue you into counter matchups, if in the game itself they are still rigging every aspect of the game as you suggested earlier? It makes zero sense. If an algorithm decides who wins as soon as the game starts, what difference would it make which class fights which?
But your convictions are not based on good data. 1,500 years ago, people were convinced that the Earth was the center of the universe. 500 years ago people were convinced that the Earth was flat. People today are convinced that the 2020 election was stolen despite absolutely no evidence to support that conclusion.
In another one of these threads, someone had linked an analysis from either VS or HSReplay where they did look at matchmaking and in the hundreds of thousands of games, found no correlation between win streaks, deck choice, or money spent and the matchups you get. Maybe someone familiar with this analysis can link it.
Again, that’s an extremely small sample size and suffers from confirmation bias.
Correct, because sample size, confirmation bias, and variable control matter.
Important note to people reading this thread is that it is (proof) to no one. No one player can see the meta changing in real time, as it’s formed from several million games.
It would be like lining up every person in your state and after seeing 5 guys in blue shirts in a row and deciding people are using mostly blue shirts. Even if you saw 50, or 500 blue shirted people you would have seen such a tiny fraction as to be meaningless information.
Well; I am satisfied with what I have seen.
I have played this game every day for eight years.
I do not believe I have missed a day.
That’s an awful lot of observation.
I’m just going off the data from HSReplay. It’s not a ton of games since they reset it not to
long ago and Poison Rouge isn’t very popular atm. But they have Poison Rogues winning 53% of their games against Control Warriors
Edit - Poison Rogues also have more games recorded against Ramp Druid than any other deck. The meta hasn’t changed for the other players.
