I would like nothing more than for people to be interested in the variables I know that play a part in why the data are never 100% right
But noone asks xD
But it looks like now at least people should be more inclined to believe that those data aren’t infallible
My understanding is that D4-D1 matchmaking is based on (hidden) player MMR. At Legend it’s based on your actual Legend rank. Please correct me if I’m mistaken.
If that is true, some of the data could be explained by tendencies/preferences of a player pool that is a mixture of ‘hard stuck at Diamond every month’ (which would face each other) and ‘just climbing to Legend as usual’ players (which would face each other even in Diamond ranks).
In my estimation that would increase the likelihood of mirrors at D4-D1 because you have two smaller populations that are more similar to each other than the overall population.
Personally I’m right on that cusp, in that I can make Legend any given month if I play enough games (or don’t grow bored), and usually switch from the D4-D1 ‘best netdeck that is the least boring to me’ to the Legend ‘best netdeck’ once I hit D4 in preparation for at least a few days in Legend ranks.
You could write a colossal list if you really wanted 
True
Yep, that’s one of the variables.
That’s also how you can see 6% of hunter players in top 1k legend, but where I play, in top 500, there are 0, which means there’s actually 12% hunters in 500-1k legend (according to data from 25-30 april) - those are probably mid-legend players who highroll a bit and get to 500-1k for a little while before returning back to where they belong
I don’t have the time and patience right now to read carefully the way you handled this data and understand it, so I can’t say if there is something wrong with it, but on a surface level read the first thing that comes to mind is that the overrepresentation of warrior mirrors could just be a sample bias.
That said, the only side of the ‘rigged’ argument that I have ever suspected to have merit in my play experience actually aligns with this finding. I have noticed some inconsistent matchup spreads when you run meta vs off-meta decks. This has made me wonder if they try to soften the ladder experience a tiny bit by making top-performing meta decks encounter each other more frequently, so junky homebrews can lose in peace vs other junky homebrews. I usually dismiss this thought as simple confirmation bias, but I do think this version of ‘rigging’ actually makes some sense from a financial perspective instead of the usual ‘bli$$ard makes me lose unless I pay for skinzz weee’.
false.
Matching without bonus stars below leged is based on rank alone.
Matching in legend is based on MMR.
They’re exactly backwards.
We’ve been told that the game uses only wins and losses to determine your ranking, but it does this by comparing your wins and losses to other players - basically looking at common games or opponents.
So by just looking at that, the meta decks are going to move into a different category over time based on what they farm and what farms them. It doesn’t even need to know any cards in your deck to know what your bad match ups are.
Ronmexico is single handedly keeping Hunter alive in top Legend on NA. When Cantelope is playing he usually gravitates back to Hunter in these types of metas as well. That’s pretty much the only players i know that play Hunter at that high of legend.
Yeah, and I’m not sure if their sample size together is enough to even be reported
They don’t play a lot - Cantelope prefers BG, and he’s stuck around 1-2k for months now, and Ron Mexico plays as slow as if he’s playing Reno Rheastrasza Druid and doesn’t stream every day + his chat makes him play whole lot of trash decks for channel points
Nah he was top 500 before the last balance patch. He just doesn’t play enough in this meta to care much. It’s not his cup of Tea. You are right though he plays a lot more BG’s. I have talked to him a bit while playing BG’s at the same time.
Players with trackers are more serious players. So they tend to play the popular / good decks more than others. While a whole onslaught of casual players isn’t counted.
I literally counted them, read the thread
Could be the problem of it being D5 and up. Perhaps one mirror is just far more popular in lower ranks or something.
Still worshipping at the altar of the stats God ? How’s that Kool Aid taste like ?
No offense intended but arguments with stats as their foundation pretty much don’t work unless it’s coming from a Blizzard employee and it’s an approved statement.
Everything else is just assuming and backing up your assumption with various unentitled reasons / perspective.
All 3rd party statistics should be taken with a grain of salt, and turning them from assumption into beliefs is just hillarious and sad.
VS data report isn’t fake, it’s just on a very small sample size, and while their data may be true for that set number of games they recorded, it’s not true for the % of the players / games that were not counted by their collectors.
60 k games is a fart in a hurricane if you do the maths of players / games played on a daily basis. So far my calculation are in the rage of 50 to 100 milion games per day, so 60 k … well its nothing.
https://live-production.wcms.abc-cdn.net.au/064a119ef5f94d92afcc4da13a9fd43b?impolicy=wcms_crop_resize&cropH=300&cropW=300&xPos=0&yPos=0&width=862&height=862
I’m not exactly sure what you’re saying you think the data is saying. What’s the conclusion?
Also, not sure if this messes with your numbers, but does the data take into account that D4-D1 is basically the “special” area for matchmaking due to how MMR matching works?
Take a player like me. I get 10 bonus stars.
From Bronze10 - D5, I’m not facing Bronze10-D5 players. I’m facing Legend MMR players that likely hit Legend last month same as I did. At no point in my run in those areas am I facing someone of equal rank unless it’s coincidence at they are also currently at that rank on their run to Legend.
Then, once D4 hits, I’m facing D4 people because I have ran out of bonus stars.
I’m not sure if that messes with the data for trackers, but I imagine it would.
If I’m at Silver 10 and fighting someone else, the data cannot assume the other person is Silver 10 because they aren’t. So is the game tracked as a Silver 10 game? What if the other person is at Diamond 10? Is that a Silver game or a Diamond game?
If my tracker assumes they are also Silver 10 and tracks it as a Silver 10 game, it would be wrong. If they have a tracker and tracks it as a Diamond 10 game, it would also be wrong. Thus, we have this overlap where my tracker submits data one way and their tracker submits data the other way, neither of which is right.
I think these overlapping areas have to be taken into account.
Since you are lazy ill do some quick math for you.
HSreplay has a live data feed that shuffles between 3 and 8 games per second ( 8 on prime hours ). Taking a lower average of 5 / second, it means 300 games / minute, 18 k / hour, 432 k games daily. Thats just standard / collectors / upper echelon.
Now, if you dont understand simple maths, thats fine, you can go on quoting memes like the average basement dweller. 
I’ve seen people be wrong about a lot of things on these forums, but that is up there with the most wrong I’ve ever seen someone be before. If you truly believe 50-100 million games are being played per day, it’s not Scr0tie that needs to understand “simple maths”, it’s you.
We’re only talking about Standard here. The context of this thread is Standard only.
Fact: if you add every mode together (BGs, Arena, Standard, Wild, etc) there hasn’t been a single day in the past month where more than 1.5 million accounts logged on, usually less. This is operating system data, so very trustworthy. So what you’re basically saying is that the average player plays 40-80 games of Standard per day. Keep in mind that many Hearthstone players don’t play Standard at all.
Awww, you didnt pass elementary school, thats cute.
Do you understand what upper echelon means ? Since you clearly don’t, and im in a gifting mood, ill simplify it for you. Most of the players of standard are those that maybe hit legend or become stuck in Diamond, and the stats that you are reading are mostly upper echelon. Theres a statistical curve that you cant see, and i think i’ve gone far enough with the explanation.
Ever heard of the Median Line ?
I’m assuming that like you’ve experienced, that D4-1 players with the tracker are rarely matched with players who aren’t D4-1 themselves. VS classified games by tracker player rank, ignoring opponent rank, so if this does happen that’s how VS counts it.
I’m not looking at D10-5 here at all.
I don’t think it’s that big of a deal. I mean, it probably is an effect that exists but the popularity of decks on the non tracker side is still the popularity of decks on the non tracker side. This kind of matchmaking doesn’t violate any fundamental assumptions in the math or anything.
On the last day of the month this is true. And the data I’m using covers April 25-30 so it’s true for the data. It wouldn’t be true today, but give it another 3 weeks and it’ll go back to being true again.
Is that because it’s invisible to everyone, or just because I’m allegedly stupid? Can you see it?
I’m looking at D4-1 data. Like you already said, a majority of players get here (by the end of the month). That’s not upper echelon, that’s median. The median player in terms of overall skill is literally in the data I’m analyzing here.
Once again: please, tell me more.