I‘m totally the opinion that there is an algorithm in place which at least pairs decks with similar winrates against each other.
Just switch your deck on ladder and accidentally you‘ll see a whole bunch of decks you never encountered with the other deck. This leads to the conclusion that not only the winrate but the winrate against certain decks (/set of cards in a deck) is considered.
Oh and as I saw the argument about Blizzard does not even bother about glitches so why should they build a complex system for matchmaking:
1st: frustrating matchups are demotivating the player base more than graphical glitches
2nd: as I‘m in the software industry, too I can tell you that the client on your smartphone may seem big, but that‘s mainly because of the many high res images. The party starts in the backend and Blizzard has already a lot of experience in algorithms due to World of Warcraft. ;)
tldr;
If Blizzard has the possibility (and they have it) to improve player experience by influencing matchmaking (with all the mentioned negative effects) why shouldn’t they use it?
to counter your theory: i was trying out midrange hunter from rank 10. after i got rolled over by odd pali, odd mage, odd rogues and spell hunters, i switched to Control Warrior. then i got matched up aginst zoo lock like 5 out 7 times.
I love how so many talk as if match making is perfectly random and this is a known fact. I haven't seen any evidence of that, and I'd love to see that instead of just smug assertions.
It's actually pretty challenging to make a truly random algorithm, much easier to approximate randomness.
Is there any incentive for Blizzard to make an algorithm that is not "truly random"? Of course, Blizzard has a bunch of goals- all centered around maximizing profit. Blizzard is a for-profit enterprise.
Motivations could include: varying matches to improve the player experience (I haven't seen much of that, but definitely possible). Queuing people into counters who are on a win-streak or just about to reach a new ranking. This could cause them to tech their deck with new cards or try a new deck, increasing the likelihood they need to craft cards or open packs. Could also be to stop players on losing streaks, as they might rage quit the game entirely. Along the same lines, it could be especially prevalent at ranks 1 and 2, players that might reach legend and lose motivation to play after achieving that key accomplishment.
I'm sure there are other motivations, but its not unreasonable to question the assertion Blizzards matchmaking algorithms are completely fair/randomized, as other parts of the game are most definitely not, for example: the appearance rates of cards in arena drafts is micromanaged in bucket systems, packs have pity timers etc.
I wish they'd just release a "Stats" in-game so we could see for ourselves what our WR is, how many games played against <insert class>, how many won, in-game time etc. instead of having to download some buggy program.
But yeah, random sometimes just doesn't feel 'random', hence why people declare miracles or some other godly-power has helped or hindered them. That's life, it's frustrating, I say you either pray more to your whatever and see if that works (it won't) or take a break, smoke a ciggy, have a wank, take a walk and come back later and try again.
I rarely see people talk about the other side of the coin. Wouldn’t that mean that others are getting matched up against consistently favored matchups?
Yeah but i don't think it's a valid argument, because firstly we don't answer to a question with a question and secondly because it doesn't answer to the first question at all. Why do we get certain specific matchups according to this or that deck ?
I guess if those threads wouldn’t pop up here and there, it would actually be a sign that the matchmaker is rigged.
But winning or losing streaks are normal.
It reminds me of that meme where the one guy presents a random generator, and he only says 9,9,9,9, etc. the one guy says, that doesn’t sound random.... ( But over a long samplesize, it’s randomly happening eventually) while the other guy says something like : you can never know, as it could be random...
And it’s very normal such streaks happen, since the average can result from spikes in both directions ( but obviously positive streaks aren’t that much represented here) .
Wonder if longterm MtG Arena, Shadowverse or Eternal players have the same discussions. From what I've been reading here a common consensus seems to be that ActiBlizz manipulate the matchmaking algorithm to make money. Not a crime in and of itself to want to make money of course, even if having a hand in matchmaking is low AF.
Not being a member of either Camp (I play to enjoy myself and could care less) this Thread has proven to be an interesting read for sure.
At this point, not a single person who claims such a thing bring.. not even evidence, just SOMETHING. At least someone who could put down "I was wondering about this so I ran a test. I played X games with this deck and tracked my games...."
At least that would offer SOME kind of discussion. NOTHING like that shows up beyond some vague "I played some 20 games or so.. or was it 50, or 5..." or just like, 3 games then that's ABSOLUTE PROOF.
That and a whole ton of theorycrafting. And conspiracies. You know, because nothing makes people want to buy more packs than making all of the cards they just opened absolutely worthless.
At this point, we've heard so many folks who SWEAR there's a conspiracy and so little actual evidence that these threads just get locked after a while.
I'm guessing if someone brought even some actual testing it may stay longer. Otherwise.. meh.
There's little point in bringing statistical evidence because most people don't understand the technical and mathematical aspects of statistical arguments particularly well. They just trot out the glib attack phrases like "confirmation bias" "small sample size" and the ever-witty "tinfoil hat" It becomes frustrating discussing stats with people who don't get the maths and trot out all kinds of irrelevant side arguments that seem logical on the surface but in fact are often not only irrelevant they are completely incorrect.
It's also the case that significant statistical evidence is usually formally defined to be the occurrence of some test statistic taking a particular value which is unlikely under some particular set of assumptions. This means that if you get a significant result, say something happens with p < 0.05, it could simply be "bad luck" -- statistical evidence is never irrefutable. For example if I assume a coin is fair and toss it 20 times and get 15 heads, then I have statistical evidence against the assumption of a fair coin (in this case the chance of seeing 15 or more heads is around 0.02) ... but it could just be a weird run of luck. Idiots would still try to argue small sample size.
Post hoc analysis of aggregate historical samples taken from a non-stationary time-varying population (like most large hearthstone datasets) can't really tell you much. Most people reading this simply will not get that. Good statistical exploration of hearthstone matchmaking would most likely require a smallish number of games played under a short time frame with randomisation aspects and coordination of multiple accounts.
I don't know for sure that ladder matchmaking is NOT completely random within your rank and NOT affected by the cards you run... but based on 17000+ ladder wins, a few goes at collecting statistical evidence, and a graduate diploma in applied statistics over the top of degree in pure mathematics, it's my strong feeling that matchmaking is manipulated in ways that Blizzard have been lying about for a long time.
I'm not asking for absolute statistical perfection or irrefutable evidence of the matter. I'm not even asking for people to abandon the notion of rigged matchmaking. I don't think we're at the point where any of that is necessary to have a serious discussion about the matchmaking system. It WILL be eventually, but not now. The first step comes from something a little more basic.
To use the example of the coin, if I were to pick up a coin and say "this coin is rigged, it keeps showing heads more than tails." I will get pushback since all I am doing is throwing a claim with no evidence. It would be like I claimed that you were a shill for Cygames, trying to belittle Blizzard's name in defense of Shadowverse. Without any actual argumentative point as to why, all we would be doing is throwing accusations at each other.
However, if I were to say "I'm worried about the coin's validity. I kept a running of 20 flips and got heads 15 times" then at least we would have SOMETHING to work with. We could at least say "that is unusual" then ask others to do similar tests. Others can then start testing with that same basis and eventually lead us to something a bit more statistical.
We're a bunch of random folks playing a game, not statistical majors, so we don't need to reach the point of statistical proof. But after enough people start showing results that show something 'interesting' we can at least have enough for a random forum community to start to question the status quo.
This HAS happened before three times in my own watching. The first was done on reddit and formed the basis of the Pity Timer and is well documented. The second occurred in the summer of 2017.
It started after the 'legendary in the first 10 packs' matter. It was declared, by Blizzard, that people who already opened a pack in a set won't be affected. However, people started reporting that they were getting a legendary within the first 10 packs anyway. The ones with vague "I opened a few and got something" were ignored, but a few others ran tests and noted their results with most showing the same 'oddness'. Even if you could declare it to be 'bad luck' it was enough to get more to test it out and post what they found and to quiet the dissidents.
Eventually it was noted that about 75% of the folks who did it got a legendary in the first 10-15 packs. Blizzard never stated anything about it, but it was generally agreed that it was NOT just RNG. For the record, I eventually tried it and got legendaries in the first 10 in every set.. except for the Classic set which gave me one after 13.
The third started a while after the second as a general "I didn't get the legendary I expected" which, of course, got push back. Then a few others noted that they track their cards and shows us the results of that tracking and how, indeed, they missed a legendary in 40 packs. Some DID argue against it, but it was enough to cause others to start collecting data. Others then joined in and, while some showed that they were within the norm, others were able to declare that things were off. It wasn't PROOF but it was enough to question whether a bug was involved.
It ended when Blizzard reported that a bug affected the draw rates of cards, proving compensation for those who didn't get the cards' they were supposed to' confirming the suspicions.
In all of those cases, we didn't have absolute statistical proof nor did we have absolute proof that no one was faking anything. But we DID have numbers to test and people willing to test them, and what started as an idea that was brushed as 'tin foil' turned into something we commoners could agree on. A scientist would've mocked us, but it was enough for the community. THAT is what I've been asking for for YEARS that I've seen this argument about matchmaking.
I have YET to see anyone even try it. I *HAVE* been fussed at for 'not undertstanding' and been accused of being a 'fan boy'. I've been talked to about theories, about what could happen, about how it MUST be true because *GREED* and *ACTIVISION* and how DARE I not believe you when the TRUTH IS SO IN FRONT OF YOU!
And I've seen plenty of "there's no point in trying, no one will believe us."
If you want to believe in something, you are free to do so. I understand about the matter of belief and consider it folly to say "only believe things that you can prove." That's stupid because you technically can't 'prove' anything. At some point, you have to have faith that something will happen without knowing fully how. Thus believing that matchmaking is rigged is a valid thing to be. But so is believing that it's NOT rigged. And so is saying "I do not believe you." And it doesn't make you a fanboy or blind for disagreeing with your beliefs. Neither does it make you a tinfoil hat wearer for believing it.
But if you decide to make a post saying "Matchmaking IS rigged." don't expect everyone to start blindly following your logic. Don't get defensive when we disagree with you, especially after YEARS of hearing it with nothing more that vague appeals to theory and unrecorded 'casual experiments' that may not even be accurate. Don't think we are simply worshiping Blizzard because we're sticking to the status quo in the face of little reason to believe otherwise. We are just following standard debate procedure: If you make a claim that changes the standard, come with evidence or we stick to the standard.
Don't be shocked when people disagree with you. Don't cry "no point in testing." when no one has tried and other theories HAVE been tested and convinced the public that the standard isn't true. Just do SOMETHING other than raging about how you're so sure and belittling those that disagree. That's ALL I ask in these threads. I have yet to get it.
I would like to sometime. Until then, you get 10 pages of essay about how to debate a topic.
I guess if those threads wouldn’t pop up here and there, it would actually be a sign that the matchmaker is rigged.
But winning or losing streaks are normal.
It reminds me of that meme where the one guy presents a random generator, and he only says 9,9,9,9, etc. the one guy says, that doesn’t sound random.... ( But over a long samplesize, it’s randomly happening eventually) while the other guy says something like : you can never know, as it could be random...
And it’s very normal such streaks happen, since the average can result from spikes in both directions ( but obviously positive streaks aren’t that much represented here) .
Man, I just read essentially "I have a graduate degree in statistics . . . my feeling is the game is rigged"
Come on, if you have a 1700 sample size (not that that would even reach a p<0.10, but whatever) at least make some kind of factual statement on that set.
I am inclined to believe you about your qualifications, if only because only people that work with statistics a lot get that jaded about what their own science can and can't say. I only have the undergraduate econometrics creds, but this is a really simple problem we're facing here. I understand what you're saying about the limitations of analysis, but assuming the data is congruent with the assumptions, you could very easily say, "I am 90%/95%/99% sure that the premise that matchmaking is not random when considering deck type". That would be proof enough for most.
The problem is, every data set that has ever been published aggregating more than 1000 games has not come anywhere near the numbers you would need to make that claim.
So what makes you say "my feeling is it's rigged" and why would you need to refer to feelings?
And on the subject of "no one here is educated enough to understand the numbers", there are quite a few folks trolling around here with enough knowledge on the subject to provide some corroboration, but again, you have to post the numbers.
Rollback Post to RevisionRollBack
Helpful Clarification on Forbidden Topics for Hearthstone Forums:
Enjoying Americans winning in the Olympics is forbidden because it is political. A 14 plus page discussion of state-sponsored lawsuits against a multi-national corporation based on harassment, discrimination, and wrongful death allegations is apparently not political enough to raise an issue.
The problem is, every data set that has ever been published aggregating more than 1000 games has not come anywhere near the numbers you would need to make that claim.
I honestly don't think we even need 1000 games for right now. at the moment we have nothing to work off of. Even something as small as 30-50, or just a person recording the games of the past month would be enough just to show that there's something we should test a little more. That is, we aren't at the point of needing to prove that it's happening, but just need to prove that further tests are needed. As I said, if 2 people flipped a coin and both got 15 out of 20, it would not prove anything, but it would be enough to ask "Hey, can we look in on this?"
Over the years of these threads, I've seen ONE person post anything interesting (though buried in the middle of a thread), a few that said they would post something then not, and a LOT of "I feel" or "I just know" or "I played 3 games, OMG RIGGGEDDD!"
For me it's clearly manipulated. And I mean everything related to RNG in this game. Just look at Trolden and the best of 2018 for example. When chances of 0.00002% occur this often, there is an issue.
It might not be intentional though. If they are using a custom made RNG (or not), then this could explain it. We all know that the "default" one, from any programming language, is not really "random" as we humans perceive it.
https://patentimages.storage.googleapis.com/f9/c9/77/b0b7cfc6883c41/US20160005270A1.pdf
The fact that this even exists is fucking atrocious
I‘m totally the opinion that there is an algorithm in place which at least pairs decks with similar winrates against each other.
Just switch your deck on ladder and accidentally you‘ll see a whole bunch of decks you never encountered with the other deck. This leads to the conclusion that not only the winrate but the winrate against certain decks (/set of cards in a deck) is considered.
Oh and as I saw the argument about Blizzard does not even bother about glitches so why should they build a complex system for matchmaking:
1st: frustrating matchups are demotivating the player base more than graphical glitches
2nd: as I‘m in the software industry, too I can tell you that the client on your smartphone may seem big, but that‘s mainly because of the many high res images. The party starts in the backend and Blizzard has already a lot of experience in algorithms due to World of Warcraft. ;)
tldr;
If Blizzard has the possibility (and they have it) to improve player experience by influencing matchmaking (with all the mentioned negative effects) why shouldn’t they use it?
Case Closed.
It is rigged af
i think ppl get tilted way too quick.
to counter your theory: i was trying out midrange hunter from rank 10. after i got rolled over by odd pali, odd mage, odd rogues and spell hunters, i switched to Control Warrior. then i got matched up aginst zoo lock like 5 out 7 times.
so i‘d say it is in fact pretty random
I love how so many talk as if match making is perfectly random and this is a known fact. I haven't seen any evidence of that, and I'd love to see that instead of just smug assertions.
It's actually pretty challenging to make a truly random algorithm, much easier to approximate randomness.
Is there any incentive for Blizzard to make an algorithm that is not "truly random"? Of course, Blizzard has a bunch of goals- all centered around maximizing profit. Blizzard is a for-profit enterprise.
Motivations could include: varying matches to improve the player experience (I haven't seen much of that, but definitely possible). Queuing people into counters who are on a win-streak or just about to reach a new ranking. This could cause them to tech their deck with new cards or try a new deck, increasing the likelihood they need to craft cards or open packs. Could also be to stop players on losing streaks, as they might rage quit the game entirely. Along the same lines, it could be especially prevalent at ranks 1 and 2, players that might reach legend and lose motivation to play after achieving that key accomplishment.
I'm sure there are other motivations, but its not unreasonable to question the assertion Blizzards matchmaking algorithms are completely fair/randomized, as other parts of the game are most definitely not, for example: the appearance rates of cards in arena drafts is micromanaged in bucket systems, packs have pity timers etc.
This is nothing to do with f2p players vs paid players.
That matchmaking always happens either you try it out on 100 games or 5 games. It happens to me a lot too.
It is simply murphy's law :)
Glad to see a civil and productive discussion on the topic. A lot of good arguments in this thread.
I wish they'd just release a "Stats" in-game so we could see for ourselves what our WR is, how many games played against <insert class>, how many won, in-game time etc. instead of having to download some buggy program.
But yeah, random sometimes just doesn't feel 'random', hence why people declare miracles or some other godly-power has helped or hindered them. That's life, it's frustrating, I say you either pray more to your whatever and see if that works (it won't) or take a break, smoke a ciggy, have a wank, take a walk and come back later and try again.
गञ्जा गञ्जा
Yeah but i don't think it's a valid argument, because firstly we don't answer to a question with a question and secondly because it doesn't answer to the first question at all. Why do we get certain specific matchups according to this or that deck ?
I guess if those threads wouldn’t pop up here and there, it would actually be a sign that the matchmaker is rigged.
But winning or losing streaks are normal.
It reminds me of that meme where the one guy presents a random generator, and he only says 9,9,9,9, etc. the one guy says, that doesn’t sound random.... ( But over a long samplesize, it’s randomly happening eventually) while the other guy says something like : you can never know, as it could be random...
And it’s very normal such streaks happen, since the average can result from spikes in both directions ( but obviously positive streaks aren’t that much represented here) .
Wonder if longterm MtG Arena, Shadowverse or Eternal players have the same discussions. From what I've been reading here a common consensus seems to be that ActiBlizz manipulate the matchmaking algorithm to make money. Not a crime in and of itself to want to make money of course, even if having a hand in matchmaking is low AF.
Not being a member of either Camp (I play to enjoy myself and could care less) this Thread has proven to be an interesting read for sure.
4/3/19 RIP Keith Flint. 😔
I'm not asking for absolute statistical perfection or irrefutable evidence of the matter. I'm not even asking for people to abandon the notion of rigged matchmaking. I don't think we're at the point where any of that is necessary to have a serious discussion about the matchmaking system. It WILL be eventually, but not now. The first step comes from something a little more basic.
To use the example of the coin, if I were to pick up a coin and say "this coin is rigged, it keeps showing heads more than tails." I will get pushback since all I am doing is throwing a claim with no evidence. It would be like I claimed that you were a shill for Cygames, trying to belittle Blizzard's name in defense of Shadowverse. Without any actual argumentative point as to why, all we would be doing is throwing accusations at each other.
However, if I were to say "I'm worried about the coin's validity. I kept a running of 20 flips and got heads 15 times" then at least we would have SOMETHING to work with. We could at least say "that is unusual" then ask others to do similar tests. Others can then start testing with that same basis and eventually lead us to something a bit more statistical.
We're a bunch of random folks playing a game, not statistical majors, so we don't need to reach the point of statistical proof. But after enough people start showing results that show something 'interesting' we can at least have enough for a random forum community to start to question the status quo.
This HAS happened before three times in my own watching. The first was done on reddit and formed the basis of the Pity Timer and is well documented. The second occurred in the summer of 2017.
It started after the 'legendary in the first 10 packs' matter. It was declared, by Blizzard, that people who already opened a pack in a set won't be affected. However, people started reporting that they were getting a legendary within the first 10 packs anyway. The ones with vague "I opened a few and got something" were ignored, but a few others ran tests and noted their results with most showing the same 'oddness'. Even if you could declare it to be 'bad luck' it was enough to get more to test it out and post what they found and to quiet the dissidents.
Eventually it was noted that about 75% of the folks who did it got a legendary in the first 10-15 packs. Blizzard never stated anything about it, but it was generally agreed that it was NOT just RNG. For the record, I eventually tried it and got legendaries in the first 10 in every set.. except for the Classic set which gave me one after 13.
The third started a while after the second as a general "I didn't get the legendary I expected" which, of course, got push back. Then a few others noted that they track their cards and shows us the results of that tracking and how, indeed, they missed a legendary in 40 packs. Some DID argue against it, but it was enough to cause others to start collecting data. Others then joined in and, while some showed that they were within the norm, others were able to declare that things were off. It wasn't PROOF but it was enough to question whether a bug was involved.
It ended when Blizzard reported that a bug affected the draw rates of cards, proving compensation for those who didn't get the cards' they were supposed to' confirming the suspicions.
In all of those cases, we didn't have absolute statistical proof nor did we have absolute proof that no one was faking anything. But we DID have numbers to test and people willing to test them, and what started as an idea that was brushed as 'tin foil' turned into something we commoners could agree on. A scientist would've mocked us, but it was enough for the community. THAT is what I've been asking for for YEARS that I've seen this argument about matchmaking.
I have YET to see anyone even try it. I *HAVE* been fussed at for 'not undertstanding' and been accused of being a 'fan boy'. I've been talked to about theories, about what could happen, about how it MUST be true because *GREED* and *ACTIVISION* and how DARE I not believe you when the TRUTH IS SO IN FRONT OF YOU!
And I've seen plenty of "there's no point in trying, no one will believe us."
If you want to believe in something, you are free to do so. I understand about the matter of belief and consider it folly to say "only believe things that you can prove." That's stupid because you technically can't 'prove' anything. At some point, you have to have faith that something will happen without knowing fully how. Thus believing that matchmaking is rigged is a valid thing to be. But so is believing that it's NOT rigged. And so is saying "I do not believe you." And it doesn't make you a fanboy or blind for disagreeing with your beliefs. Neither does it make you a tinfoil hat wearer for believing it.
But if you decide to make a post saying "Matchmaking IS rigged." don't expect everyone to start blindly following your logic. Don't get defensive when we disagree with you, especially after YEARS of hearing it with nothing more that vague appeals to theory and unrecorded 'casual experiments' that may not even be accurate. Don't think we are simply worshiping Blizzard because we're sticking to the status quo in the face of little reason to believe otherwise. We are just following standard debate procedure: If you make a claim that changes the standard, come with evidence or we stick to the standard.
Don't be shocked when people disagree with you. Don't cry "no point in testing." when no one has tried and other theories HAVE been tested and convinced the public that the standard isn't true. Just do SOMETHING other than raging about how you're so sure and belittling those that disagree. That's ALL I ask in these threads. I have yet to get it.
I would like to sometime. Until then, you get 10 pages of essay about how to debate a topic.
One does not simply walk into Mordor,
unless they want to be the best they can be.
https://dilbert.com/strip/2001-10-25
Man, I just read essentially "I have a graduate degree in statistics . . . my feeling is the game is rigged"
Come on, if you have a 1700 sample size (not that that would even reach a p<0.10, but whatever) at least make some kind of factual statement on that set.
I am inclined to believe you about your qualifications, if only because only people that work with statistics a lot get that jaded about what their own science can and can't say. I only have the undergraduate econometrics creds, but this is a really simple problem we're facing here. I understand what you're saying about the limitations of analysis, but assuming the data is congruent with the assumptions, you could very easily say, "I am 90%/95%/99% sure that the premise that matchmaking is not random when considering deck type". That would be proof enough for most.
The problem is, every data set that has ever been published aggregating more than 1000 games has not come anywhere near the numbers you would need to make that claim.
So what makes you say "my feeling is it's rigged" and why would you need to refer to feelings?
And on the subject of "no one here is educated enough to understand the numbers", there are quite a few folks trolling around here with enough knowledge on the subject to provide some corroboration, but again, you have to post the numbers.
Helpful Clarification on Forbidden Topics for Hearthstone Forums:
Enjoying Americans winning in the Olympics is forbidden because it is political. A 14 plus page discussion of state-sponsored lawsuits against a multi-national corporation based on harassment, discrimination, and wrongful death allegations is apparently not political enough to raise an issue.
There is no big mistery here to be solved, it is just confirmation bias with a small sample, nothing more.
A well crafted sharpshooter fallacy is much better challenge for skeptical people, what Shadowrisen probably is looking for. :P
I honestly don't think we even need 1000 games for right now. at the moment we have nothing to work off of. Even something as small as 30-50, or just a person recording the games of the past month would be enough just to show that there's something we should test a little more. That is, we aren't at the point of needing to prove that it's happening, but just need to prove that further tests are needed. As I said, if 2 people flipped a coin and both got 15 out of 20, it would not prove anything, but it would be enough to ask "Hey, can we look in on this?"
Over the years of these threads, I've seen ONE person post anything interesting (though buried in the middle of a thread), a few that said they would post something then not, and a LOT of "I feel" or "I just know" or "I played 3 games, OMG RIGGGEDDD!"
One does not simply walk into Mordor,
unless they want to be the best they can be.
For me it's clearly manipulated. And I mean everything related to RNG in this game. Just look at Trolden and the best of 2018 for example. When chances of 0.00002% occur this often, there is an issue.
It might not be intentional though. If they are using a custom made RNG (or not), then this could explain it. We all know that the "default" one, from any programming language, is not really "random" as we humans perceive it.
It needs to be made more "imperfect".
Show some actual statistics or it didn't happen.
Without a large data set, this thread is pointless.
Our dumb monkey brains are designed to jump to conclusions with small data sets.
This, and similar threads have turned very poisonous, so we do not allow them on this site. Locked.
Editor of the Heartpwn Legendary Crafting Guide:
https://www.hearthpwn.com/forums/hearthstone-general/card-discussion/205920-legendary-tier-list-crafting-guide