Suspicious Matchmaking

Submitted 4 years, 3 months ago by

I have been playing a lot of Galakrond Warlock, but I noticed, that most of my opponents were Shamans. Against those I really struggled to win, so i looked up, which deck was best against Shaman. Apparently Holy-Wrath-Paladin is doing pretty well against them and since I had all the necessary cards, I thought I'd give it a try. I played 10 games with it, but not a single Shaman (mostly Rogue, which is the strongest deck against Holy-Wrath-Paladin :D).

Then I switched back to Warlock and I faced 3 Shamans in a row, after that I stopped playing. The last SEVEN games with my Warlock Deck were against Shaman.

I'm not salty, I have a positive winrate after all, I just find it suspiciously unexpectable (statistically speaking). Has someone of you encountered a similar situation?

  • Mephisto's Avatar
    185 19 Posts Joined 12/15/2019
    Posted 4 years, 3 months ago

    I have been playing a lot of Galakrond Warlock, but I noticed, that most of my opponents were Shamans. Against those I really struggled to win, so i looked up, which deck was best against Shaman. Apparently Holy-Wrath-Paladin is doing pretty well against them and since I had all the necessary cards, I thought I'd give it a try. I played 10 games with it, but not a single Shaman (mostly Rogue, which is the strongest deck against Holy-Wrath-Paladin :D).

    Then I switched back to Warlock and I faced 3 Shamans in a row, after that I stopped playing. The last SEVEN games with my Warlock Deck were against Shaman.

    I'm not salty, I have a positive winrate after all, I just find it suspiciously unexpectable (statistically speaking). Has someone of you encountered a similar situation?

    "Monsters behind you!"

    1
  • Mephisto's Avatar
    185 19 Posts Joined 12/15/2019
    Posted 4 years, 3 months ago

    I don't really know how to post an image, so here is a link to a screenshot of the decktracker:

    https://ibb.co/0tPP0r3

    "Monsters behind you!"

    1
  • Avalon's Avatar
    Salty Dog 1550 2105 Posts Joined 06/12/2019
    Posted 4 years, 3 months ago

    To be fair I meet a shit ton of Shamans no matter what I play.
    Remember that 10 (even 20) games aren't a big enough sample to get conclusions from.

    Spice Lord and self-proclaimed Meme Master.

    • You can follow me on Twitter - I am always active and you can tag me to highlight your (or someone else's) 12 wins Duels run or really anything Hearthstone-related!
    • Hearthstone Battletag: beppe946#2807 (EU)
    0
  • Mephisto's Avatar
    185 19 Posts Joined 12/15/2019
    Posted 4 years, 3 months ago

    Assuming 30% of decks are shaman decks, then the probability of not playing against shaman for 10 games in a row is less than 3%. (0.7^10)

    Pretty low, but not impossible, I know, just found it really weird, that when I switched back I only faced shamans again.

    "Monsters behind you!"

    -1
  • Alfi's Avatar
    Devoted Academic 1790 1375 Posts Joined 05/29/2019
    Posted 4 years, 3 months ago

    Basically what you are saying that Blizzard is rigging random players to have the best matchup available against you.

    -=alfi=-

    12
  • Mercynary9's Avatar
    355 239 Posts Joined 03/31/2019
    Posted 4 years, 3 months ago

    I had this happened to me too alot when I was playing on Gadgetzan expansion, theres one certain deck I faced often and lost many games, switched to another deck, and never seen that obnoxious deck again and lost several times again.

    Feels bad man when you facing rigged matchups right now, but during rise of shadow expansion I had good matchups several games. It may be suspicious right now, but you can still try to spam it until you have finally to see the opponent you want.

    But of course... taking a break is my best answer for now, I would give up at that moment honestly.

    -4
  • Mephisto's Avatar
    185 19 Posts Joined 12/15/2019
    Posted 4 years, 3 months ago

    I know I sound like some flat-earther now :D

    but yes, it seemed rigged.

    Do you think Blizzard would actually do something like that? For example to keep winrates of certain decks in control.

    "Monsters behind you!"

    0
  • Horus's Avatar
    Detective Pikachu 2565 3339 Posts Joined 05/29/2019
    Posted 4 years, 3 months ago

    Its the classic rock-paper-scissor :
    Whenever I play rock, my opponent has the paper. But when I use paper, I'm being counter with scissor ...

    Struggle with Heroic Galakrond's Awakening? I got your back : 

    2
  • CursedParrot's Avatar
    640 720 Posts Joined 05/29/2019
    Posted 4 years, 3 months ago

    No. It would be incredibly difficult for Hearthstone to implement an algorithm to target specific players, and they have no incentive to do so. However, I think it could be cool if Hearthstone implemented a system where it tried not to match you against the same class twice in a row (kind of like how shuffle on iPods was changed to be less random so songs of the same style wouldn't follow one another as much, even though it was purely random before)

    7
  • CursedParrot's Avatar
    640 720 Posts Joined 05/29/2019
    Posted 4 years, 3 months ago

    This would also make it so that someone playing a popular class would take longer to find a match than someone playing an unpopular one. It would also help counter the problem of one deck becoming a large percentage of what you face

    1
  • OmarComing's Avatar
    790 530 Posts Joined 05/28/2019
    Posted 4 years, 3 months ago

    We can’t go on together with suspicious matchmaking (with suspicious matchmaking).

    And we can’t build our decks, with suspicious matchmaking. 

    1
  • DelkoHS's Avatar
    Child of Galakrond 485 481 Posts Joined 05/28/2019
    Posted 4 years, 3 months ago

    Funny, I was also playing Holy Wrath Paladin earlier because I faced a lot of Shamans, and then I got 6 Rogues in a row (which is super annoying due to Flik Skyshiv). In 20 games I faced 1 Shaman, heck I even faced more Control Warriors..

    0
  • OldenGolden's Avatar
    Snow-Covered 690 131 Posts Joined 05/28/2019
    Posted 4 years, 3 months ago

    I'm completely a tinfoil hat guy on this subject. . . perhaps even moreso when I play MtGA. I just KNOW these SOBs are against me somehow.

    Also, unrelated - I love Galakrond Warlock, despite all the pre-release trashing :)

    -1
  • Avalon's Avatar
    Salty Dog 1550 2105 Posts Joined 06/12/2019
    Posted 4 years, 3 months ago

    Yeah I agree with you: Flik Skyshiv is one hell of a hard counter.

    Luckily for me I've always been to the offender's side until now :)

    Spice Lord and self-proclaimed Meme Master.

    • You can follow me on Twitter - I am always active and you can tag me to highlight your (or someone else's) 12 wins Duels run or really anything Hearthstone-related!
    • Hearthstone Battletag: beppe946#2807 (EU)
    0
  • Marega's Avatar
    620 872 Posts Joined 05/28/2019
    Posted 4 years, 3 months ago

    First time?

    with all honesty i thought like that a year ago. ive now adopted a different mindset.

    so the secret to fight that feeling is just stick with one deck. odds are u will be facing shaman the most so no matter what keep playing that holy wrath paladin and things will even out

    1
  • EpicHercules's Avatar
    Jaina 270 59 Posts Joined 06/21/2019
    Posted 4 years, 3 months ago

    I've experienced this on so many occasions that I lost count. I'd be playing a deck and keep running into the same class that was punishing me mercilessly, so like a sensible fellow, I'd switch decks. BOOM! Wouldn't see another at all. Like the opponents I was playing disappeared right when I switched to a counter. Don't know what it is, but I seriously don't think Blizzard targets individual players. I'm sure they use your deck as a factor in matchmaking, but how or why is a bit of a mystery.

    Full tinfoil hat time though: I have suspected in the past they do this as part of a behavioral engineering algorithm; they simply match a person they know will buy packs against counters until they get frustrated enough to buy more cards in an effort to win again. Probably complete B.S., but sounds like something I'd do if I was in charge. Lol

    0
  • KANSAS's Avatar
    Old God Fanatic 1745 2912 Posts Joined 03/25/2019
    Posted 4 years, 3 months ago

    This is a fairly silly claim. Some people forget that you are fighting another real person playing hearthstone. Whenever you get a really bad matchup, someone else gets a really good matchup. Which means that if you were unlucky and got seven bad matches in a row, there are seven people who got a really good matchup. Do you really think blizzard is randomly choosing half of their players to get bad matchups, and half of their players to get good matchups?

    Carrion, my wayward grub.

    5
  • Gerix55's Avatar
    120 29 Posts Joined 05/29/2019
    Posted 4 years, 3 months ago

    I feel it since switched to wild in ROS.

    0
  • Meteorite12's Avatar
    670 696 Posts Joined 05/29/2019
    Posted 4 years, 3 months ago

    I decided to start grinding up the ranks with my midrange hunter and I haven’t really had any problems with seeing counters. The decks that counter me are pretty evident too, anything that floods the board a lot is a likely loss for me (Don’t have anywhere near the board clears to beat something like the Warlock Galakrond or really aggressive Pirate/Murloc decks). I’ve seen plenty of different types of decks along the way and not many of them have been that bad of a matchup.

    Even in the past I don’t remember any times it felt like something was up, my good and bad matchups were generally split enough that I wouldn’t rank up, but I wouldn’t go down in ranks much either

    Who needs consistency when you could have fun?

    0
  • og0's Avatar
    Red Riding Hood 1570 1062 Posts Joined 03/31/2019
    Posted 4 years, 3 months ago

    Officially stated by Team 5.

    From https://www.hearthstonetopdecks.com/matchmaking-works-tyrandespecial-promotions-max-mccall-yong-woo/

    (the bold is my emphasis)

    MAX MCCALL DISCUSSES HOW MATCHMAKING WORKS

    I saw this post last week, but didn’t have time to respond to it. I’ll do that now: when you go into a game, the only variable that affects who your opponent will be is your skill rating.

    Matchmaking works as follows:

    We use a formula to assess player skill. After every game, the formula looks at if you won or lost and uses your current rating, your opponent’s rating, and your rating history to generate your new rating. We call this rating MMR for short. In casual and at Legend rank, we pair players with similar MMRs. In Ranked below legend, we pair people with similar star ranks instead of similar MMRs. Your rating is the only input that the matchmaker receives. It doesn’t know what deck you’re playing, what deck you just played with or against, or anything else, except for your rating.

    When you press ‘play’ you enter a queue for your chosen game mode. The matchmaker looks at your MMR and compares it to the MMR of everyone else in the queue. If it finds someone else with the same MMR as you, it pairs you into a game. If it doesn’t, it will wait a few seconds and look again. The second time, it doesn’t look just for someone with your MMR; it will also look for someone with an MMR that’s almost the same as yours. If it still doesn’t find a match, it waits another few seconds and looks again. The bound for what MMRs are considered a good match keep widening the longer you’re in the queue; this is to ensure that you don’t have to wait too long to play. Usually a match is found so quickly that the widening bounds never really matter. After the game, your rating is updated, and the process is repeated the next time you queue up.

     

    NOTE:  the above was written before they changed it to make you less likely to face the same opponent twice in a row, so that's now a factor in the match making.

    I realize some won't accept this and are convinced the game is rigged.  I'll listen if an analysis of 100K+ matchups by a statistics collection site by unbiased competent people show there is statistically significant matchup bias.   It's too easy as humans to suffer from comfirmation bias especially with a small sample size.

    For the record I fell out with Blizzard during WoW days and told them I'd never give them money ever again (I haven't).  So I wouldn't consider myself part of their fanbase :)

    All generalizations are false.

    5
  • sinti's Avatar
    Senior Writer Chocolate Cake 2070 2774 Posts Joined 10/20/2018
    Posted 4 years, 3 months ago
    Quote From Mephisto

    I know I sound like some flat-earther now :D

    but yes, it seemed rigged.

    Do you think Blizzard would actually do something like that? For example to keep winrates of certain decks in control.

    No, the consequences, if proven they are doing this, would most certainly not be worth the "upside".

    I get it, we've all had these things happen to us and variance can make it feel rigged, but like others said, few dozen games is such a small sample size, you cant draw any conclusions from that. Play thousand, tens of thousands of games, then you might start having a sample size of some relevance ... which would still be affected by other factors outside your scope of influence, not even talking about anything Blizzard related.

    Just vent a little and move on, not much else to do here :)

    ~ Have an idea? Found a bug? Let us know! ~
    ~ Join us on Discord ~

    1
  • Bersak's Avatar
    Magma Rager 720 432 Posts Joined 06/12/2019
    Posted 4 years, 3 months ago

    There is a psychological term for this.

    I don‘t remember it but i‘m certain there is one :P

    People have a tendency of percieving bad luck more frequently even if the odds are 50/50. 

    Best example: USB sticks! You allways attempt to plugg them in upside down. Or at least, it feels like it.

     

    Winner winner chicken dinner

    1
  • RandomGuy's Avatar
    430 614 Posts Joined 05/29/2019
    Posted 4 years, 3 months ago
    Quote From Mephisto

    I know I sound like some flat-earther now :D

    but yes, it seemed rigged.

    Do you think Blizzard would actually do something like that? For example to keep winrates of certain decks in control.

    Occam's Razor.

    Is it bad luck? Or is there a secret algorithm designed to pit only you against your worst match up.

    Keep in mind, that would mean that there's an algorithm designed to give someone else all of their best match ups.

    What would be the purpose of such a system?
    Why is it only giving you bad matches, but other people good ones?
    Why would resources be dedicated to design and implement that system?

    1
  • Xarkkal's Avatar
    Servant of Illidan 910 1321 Posts Joined 03/29/2019
    Posted 4 years, 3 months ago

    Holy Wrath is only good against Shaman. It's honestly not worth playing to try and counter the Shamans, because you'll just get beaten by everything else. I suggest either join the dark side (Shaman) until nerfs hit, or go play some random deck in Wild to pass the time. 

    0
  • Mephisto's Avatar
    185 19 Posts Joined 12/15/2019
    Posted 4 years, 3 months ago

    That's not true, until now I played 13 games with it and won 9 of them, without facing any shamans.

    "Monsters behind you!"

    0
  • LilianVoss's Avatar
    225 30 Posts Joined 07/20/2019
    Posted 4 years, 3 months ago

    It has been stated so far more than once by Hearthstone developers that there is a algorithm built in matchmaker that tries to keep all players around 50% win rate.

    The reasoning behind it is that players of lower skill level need help to win some matches. Winning games makes players feel good about themselves and so they keep playing and paying for the game. If "bad" players would lose constantly, they'll eventually rage quit the game and most likely wouldn't ever return to it. Every lost player is potentially one lost paying customer, so every "free to play" game out there has a variety of ways to try to keep as many players playing it as possible.

    By playing highly polarized deck like OTK Paladin, you're only making it easy for matchmaker to try to fit you within that 50% "rule". If you want to climb more reliably, turn to the dark side and play Galakrond Shaman and you'll have a much easier time.

     

    In death, I exact my revenge!

    0
  • og0's Avatar
    Red Riding Hood 1570 1062 Posts Joined 03/31/2019
    Posted 4 years, 3 months ago
    Quote From LilianVoss

    It has been stated so far more than once by Hearthstone developers that there is a algorithm built in matchmaker that tries to keep all players around 50% win rate.

    Do you have a source for that?  In my post earlier in the thread they've been clear there isn't. 

    Any 50% winrate goal would be based on the idea you rank/MMR tried to pit you against players of your skill level (or deck power level at least) over time.  Perhaps that is what was referred to.

    All generalizations are false.

    1
  • LilianVoss's Avatar
    225 30 Posts Joined 07/20/2019
    Posted 4 years, 3 months ago
    Quote From og0
    Quote From LilianVoss

    It has been stated so far more than once by Hearthstone developers that there is a algorithm built in matchmaker that tries to keep all players around 50% win rate.

    Do you have a source for that?  In my post earlier in the thread they've been clear there isn't. 

    Any 50% winrate goal would be based on the idea you rank/MMR tried to pit you against players of your skill level (or deck power level at least) over time.  Perhaps that is what was referred to.

    That's also possible. Oddly, I can't find any source atm (will try some more later), but I remember clearly reading about it more than once.

    Also, this video kinda explains what I'm talking about here. Interesting part starts at 11:07

    In death, I exact my revenge!

    1
  • og0's Avatar
    Red Riding Hood 1570 1062 Posts Joined 03/31/2019
    Posted 4 years, 3 months ago

    Yeah I do think he means the ladder system as a whole here so legend #1 doesn't play a rank 50 guy/gal.

    I've never read they try to fiddle with a same rank matchup to ensure you get a 50% winrate.  I've been able to get 85-90% winrate over 15 or so games sometimes at lower ranks.

    Again, they've stated the exact opposite which I believe in the absence of proof to the contrary.  The ladder system and meta environment as a whole acts to somewhat balance the matchups/winrates. 

    I do understand though sometimes it doesn't *feel* that way :)

    All generalizations are false.

    0
  • LilianVoss's Avatar
    225 30 Posts Joined 07/20/2019
    Posted 4 years, 3 months ago
    Quote From og0

    Yeah I do think he means the ladder system as a whole here so legend #1 doesn't play a rank 50 guy/gal.

    I've never read they try to fiddle with a same rank matchup to ensure you get a 50% winrate.  I've been able to get 85-90% winrate over 15 or so games sometimes at lower ranks.

    Again, they've stated the exact opposite which I believe in the absence of proof to the contrary.  The ladder system and meta environment as a whole acts to somewhat balance the matchups/winrates. 

    I do understand though sometimes it doesn't *feel* that way :)

    I don't think he was talking about ladder, rather about general winrate. It's simply a sound business decision for any "free to play" game to try to keep as many players as possible to play your game and, as Ben Brode in the video said, you can be sure that if your low skill players keep losing matches, they'll eventually quit.

    Also, winrate in 15 games is completely irrelevant. If you could keep 90% winrate across 150 matches, then that would be whole another story. But you can't. No one can.

    But since I still can't provide any HS developer statement that Hearthstone matchmaker works the way I described it, therefore I have no evidence so I must admit I was wrong.

    Then again, you or anyone else outside Blizzard can't prove that what Max McCall said in that post is truth either. I guess, in the end, all we have is the way we *feel* about it ;)

     

    In death, I exact my revenge!

    0
  • og0's Avatar
    Red Riding Hood 1570 1062 Posts Joined 03/31/2019
    Posted 4 years, 3 months ago

    Might not have phrased that well - I mean ladder/MMR is the Hearthstone implementation of what he was saying, but of course could be implemented differently in other games.

    I think any matchup bias at work could be determined across millions of games clearly, and of course such stats are available to some outside of Team 5/Blizzard so would be risky to do.  I also think there's not a good reason for them to do it, especially if it could be discovered and revealed.

    In my view people just suck (me included) at spotting statistically significant patterns of this type especially since HS is often an emotive gaming experience.  

    All generalizations are false.

    1
  • Thonson's Avatar
    HearthStationeer 1740 1714 Posts Joined 03/24/2019
    Posted 4 years, 3 months ago

    I wouldn't be surprised if there is a "deck MMR" or something too.  In other words, you build a deck and win a lot of games, the game sees that deck as good.  It then tries to match you up against players with similar MMR that are also playing a deck with a higher winrate.  Hence you were playing Zoolock, doing well, then you get hit with a bunch of Shaman opponents who are also winning a lot.

    You build a brand new deck of HW Paladin, the game does not know yet how this deck will perform when you are piloting it.  It matches you up against players that haven't been winning a lot, those Rogue players you mentioned.  So until you show the game that you can win a lot with this deck, you may not see the Shaman onslaught you are targeting.

    But honestly, I have no idea.

    In other words, if this rambling of mine is even slightly correct: it's an attempt to balance player experiences, not a grand plot against you.

    Quick!  Someone give me something clever to write here.

    0
  • jaypee's Avatar
    Uther 70 5 Posts Joined 07/26/2020
    Posted 3 years, 8 months ago

    Dont put your tinfoil hats down justttt yet.  lol. 

    I felt very similar to OP for a long time once I learned a few decks and started winning.  I figured it was just me being salty as many of you were telling OP.  Confirmation bias, etc etc.  I get it.  Of course I'm mad I lost those heart-breaker games, but I brushed it off as you all did OP for those same reasons.  I like to consider myself a rational man.  Anyhow... 

    These two decks constitute at least half of my all time games played.  I attached some pics below.  The 54% winrate is a Tempo DH and the 60% is a Galahrond WL. 

    If you want conclusive proof with hundreds of thousands of samples USDA certified by several PhD's in meat-statistics just so that you can consider the possibility of what OP was talking about, do not bother reading any further lol. 

    Notice:

    (1) On the Tempo DH the MOST matches have been against the worst class match-up for my specific deck by far (per the HSPREPLAY.NET global stats).  

    (2) On the Galakrond WL the MOST matches have been against the worst class match-up for my specific deck by far (per the HSPREPLAY.NET global stats).  

    (3) Look at the proportion of matches I get for the sub 50% winrate match-ups, compared to the higher winrate matchups, on both decks. 

    Regardless of how they are calculating the MMR or how they are making it so that "most people hover around a 50% winrate", I think its total bull for the simple fact that players assume they are queuing up for a random opponent of similar skill not be penalized with their worst match-ups. 

    I would love to hear from anyone else who has this data for their most played decks.  Post a screenshot of the games played breakdown and the match-up snippet up top.  I'm dying to see others' numbers. 

    EDIT: to provide actual links 

    Tempo BH Deck

    https://imgur.com/a/uPQShMY

    Galakrond WL

    https://imgur.com/a/r3GAoSg

    0
  • dapperdog's Avatar
    Dragon Scholar 1890 5543 Posts Joined 07/29/2019
    Posted 3 years, 8 months ago

    An easy enough response for anyone with suspicions about their match making.

    If everyone is paired with the worst matchup they can think off, then there is by definition no worst match since everyone's matches are equally good/bad. So there must be some pairings with good matchups. In other words, we are accusing blizzard of discrimination.

    Okay, put that aside. Let's assume blizz has control over the MMR directly, as in they can somehow match their favorite decks against its best matchups, and decks they hate into worst matchups, 80% of the time. How does blizz determine what decks fit the description? By certain cards? So if I include most of totem shaman's cards in my deck like Splitting Axe, does that screw their data sufficient enough for me to gum up their plans? If that's so then hsreplay should show some weird deck that should theoretically never work being tier 1 instead of something that makes sense.

    Lets go even further with the idea that blizz wants to maintain a near 50% win rate for everyone no matter what their deck. That's never in doubt. Blizz would prefer a more even spread instead of an odd 80% win rate deck, that's why they do nerfs. If MRR can do the same, why bother with nerfs? Just tank them with controlled MMR. Why bother with so much math, when you can simply change, introduce, or destroy decks by adjusting mana costs, or printing key cards?

    Card games are inherently filled with RNG. You can just as easily win good matchups and bad matchups provided that you draw the cards you need and your opponent doen't. You can also help up your bad matchups through deck building, and understanding your deck vs your opponent's deck. One common example is whether to play Bluegill Warrior in libram paladin or not. Simple inclusions like this can make or break the deck's win rate against priest, mage and rogue, traditionally poor matchups.

    If you let this superstition of fixed MMR take over you, then its better to just play the standard basic decks created by blizz, since you're pretty much helpless against their algorithms anyway.

     

    Small edit: this post is not a response against any specific individual. I used the word 'you' in this post as a general reference to anyone reading this.

    1
  • jaypee's Avatar
    Uther 70 5 Posts Joined 07/26/2020
    Posted 3 years, 8 months ago

    It's not superstition if there are numbers corroborating the claim.  I didnt think anything of it until I reviewed the match history of my most played decks and noticed an outsized amount of my matches were against specifically the sub 50% winrate matchups for the deck in play.  And on each deck shown, the most matches were against the specifically worst matchup. 

    You're implying that this would be impossible with the arguments you presented but none of them actually show how it's logically impossible.  

    0
  • dapperdog's Avatar
    Dragon Scholar 1890 5543 Posts Joined 07/29/2019
    Posted 3 years, 8 months ago
    Quote From jaypee

    It's not superstition if there are numbers corroborating the claim. 

    Im not trying to be a dick or anything, but unless those numbers are pried from blizz's internals or at least allegedly come from an insider source I'm prepared to challenge any one person's numbers by pointing out at the obvious; no one or two person's numbers can prove anything in this vast ocean of players between three continents. Its just as easy for me to show my pack opening percentages and try to 'prove' that I'm being jinxed because I'm hoarding more than 50k dust. Its all ultimately speculation, and from your earlier post I think you understand that.

    Quote From jaypee
    You're implying that this would be impossible with the arguments you presented but none of them actually show how it's logically impossible.

    The first and second part, or specifically the second and third paragraph, I tried my hand at logically thinking about it. If it doesn't convince you or anyone else, then that's fine. I look forward to someone else with a better mind to think up something else.

    Anyway, if its worth anything, my fifth and sixth paragraph sums it up nicely what I really think about the whole affair.

    1
  • jaypee's Avatar
    Uther 70 5 Posts Joined 07/26/2020
    Posted 3 years, 8 months ago
    Quote From dapperdog

    If everyone is paired with the worst matchup they can think off, then there is by definition no worst match since everyone's matches are equally good/bad. So there must be some pairings with good matchups. In other words, we are accusing blizzard of discrimination.

    You wouldn't need to give everyone the worst possible match-up. Simply higher odds of a bad match-up.  They have stated before that if your low-skill players get rolled every game they will stop playing (which is something for them to avoid obviously).  In my mind I thought it was just that if you lose enough games you'll start getting biased towards better match-ups, and vise-versa.  This still means low-skill players will have low win-rates and high-skill players will have higher ones, it would just have the effect of keeping both "closer" to 50% than they would be otherwise. 

    Quote From dapperdog

    Okay, put that aside. Let's assume blizz has control over the MMR directly, as in they can somehow match their favorite decks against its best matchups, and decks they hate into worst matchups, 80% of the time. How does blizz determine what decks fit the description? By certain cards? So if I include most of totem shaman's cards in my deck like Splitting Axe, does that screw their data sufficient enough for me to gum up their plans? If that's so then hsreplay should show some weird deck that should theoretically never work being tier 1 instead of something that makes sense.

    I would not assume they have "favorite" or "hated" decks.  Moreover, they would only do something like this (non truly random matchmaking) in the name of a better playing experience on average.  It would be crazy to assume they had "something against" specific decks or players-- I agree with that.  But with the amount of data (of X vs Y deck) they have, even with the "incomplete" HSreplay global summary data, you can code this into matchmaking in a way like I mentioned above.  It would not guarantee you lose or win certain games, just simply "influence" the outcome on average.  For example, its not easy to win a 42% win-rate match-up, but it is hardly impossible. 

    Quote From dapperdog

    Lets go even further with the idea that blizz wants to maintain a near 50% win rate for everyone no matter what their deck. That's never in doubt. Blizz would prefer a more even spread instead of an odd 80% win rate deck, that's why they do nerfs. If MRR can do the same, why bother with nerfs? Just tank them with controlled MMR. Why bother with so much math, when you can simply change, introduce, or destroy decks by adjusting mana costs, or printing key cards?

    To be honest, I'm not a programming or statistics buff (know a little bit though!) but its really not that difficult.  Its all just numbers and code.  Moreover, implementing an ELO/MMR type system requires a stats person.  They certainly have the expertise, resources, knowledge to create whatever type of matchmaking system they want-- its not a hassle for them.  A broken matchmaking will kill a game, so you can assume they certainly pay attention to this detail.  Particularly with an emphasis on players having a good experience (on average).  Nerfs deal with changing and adjusting the meta but there is good reason for them both to exist.  

    Quote From dapperdog

    Card games are inherently filled with RNG. You can just as easily win good matchups and bad matchups provided that you draw the cards you need and your opponent doen't. You can also help up your bad matchups through deck building, and understanding your deck vs your opponent's deck. One common example is whether to play Bluegill Warrior in libram paladin or not. Simple inclusions like this can make or break the deck's win rate against priest, mage and rogue, traditionally poor matchups. 

    You don't need to guarantee someone will lose any specific game. I.e. did they have bluegill warrior or not. All you need to do is pair them vs a 42% win-rate match-up and they will lose 42% of the time over time.  Also, changing out Bluegill Warrior will mean you are matchmaking with a different deck and there will be different data as to the worst match-ups against that specific deck (with or without Bluegill Warrior ). 

    Quote From dapperdog

    Small edit: this post is not a response against any specific individual. I used the word 'you' in this post as a general reference to anyone reading this.

    Thank you for your cordial response btw.  I hope you do not take what I was saying the wrong way.   I really do get where you are coming from, but when I saw my previous matches data it suddenly became very difficult to accept the type of rationale I was operating under before (i.e. some of the things yourself and several others have mentioned).  So I seriously get what you all are saying, but how I see it now is: 

    (1) I gots data and it dont look good.  Small as it may be. 

    (2) There is an incentive for this to be the case both from a profitability and player experience perspective.  (It also would not be dramatic enough to take good players below 50% winrate)

    (3) Many arguments around the net saying this is impossible break down into things like "but it can't be!", "too much risk to blizz to do this", "too little data". 

     

    Again, I would love to see the same screenshots I posted for anyone else.  It would be interesting to see if players' sub-50% win-rate decks are disproportionately being matched with their best matchups.  

     

    EDIT: Removed "and nobody implied that" because it was not necessary.  And made the 2nd paragraph more readable but did not change the contents. 

    0
  • LyraSilvertongue's Avatar
    360 383 Posts Joined 06/01/2019
    Posted 3 years, 8 months ago

    There's no reason for T5 to rig games.

    HS is a game that really has a poor 'end game' and poor pvp incentives. The end game for many competitive players is to get to legend which at best gets you a little more dust than lower ranks (via more packs & higher rarity cards to dust). This isn't WoW where you get unique transmog gear, titles, mounts, and FoS. Are we really to believe T5 is rigging games so players don't get a few extra hundred dust each month?

    More importantly, the common argument of T5 allegedly rigging games to get players to buy more packs is super flawed and risky for them. For example, say T5 is rigging games for players to spend more money, this is making two major assumptions. The first assumption is that T5 would be believing most players would in fact buy a lot more packs to fill their collection. The second assumption is that players wouldn't actually be the type of players that would instead say "Screw this game! If I'm going to keep losing and not have fun then I'm just going to quit this game." An algorithm can't predict which of these two players you would be and it would seem like a financially reckless gamble to sneakily make more money this way.

    A bonus hole to punch through this theory is why would T5 allegedly rig games of players who have filled the holes in their collection for the decks they care about? If you no longer have any incentive to buy more packs from a particular set of packs, until the new expansion, having your games rigged isn't going to suddenly make you want to do something (ie more pack buying) that you don't care about.

    -1
  • jaypee's Avatar
    Uther 70 5 Posts Joined 07/26/2020
    Posted 3 years, 8 months ago
    Quote From LyraSilvertongue

    This isn't WoW where you get unique transmog gear, titles, mounts, and FoS. Are we really to believe T5 is rigging games so players don't get a few extra hundred dust each month?

    I don't believe anybody claimed it was about a few hundred dust.  Just 'sayin.   I was just pointing out that OP is absolutely NOT CRAZY because instead of recounting one night of "feeling like he got a ton of bad match-ups" I have some actual match data which tells the same story.  I'm sure he was feeling like he was crazy but if he has an experience similar to mine, he's not crazy.  If he has had an experience similar to mine he's literally getting a disproportionately high amount bad match-ups.  Not saying anything more or anything less. 

    Quote From LyraSilvertongue

    The first assumption is that T5 would be believing most players would in fact buy a lot more packs to fill their collection. The second assumption is that players wouldn't actually be the type of players that would instead say "Screw this game! If I'm going to keep losing and not have fun then I'm just going to quit this game."

    If these assumptions were correct there wouldn't be any pay-to-win games, where players pay to "fill in the gaps" in those games.   But its still a very profitable type of game even with the players who play and eventually do say "Screw this game! etc etc". (To be clear, I'm not talking about HS in that example.) Moreover, OP and myself have positive win-rates in the 55-60% range, so its really not just salt when we "keep losing" lol.  It's a fun game either way, that's why we all play. 

    Quote From LyraSilvertongue

    A bonus hole to punch through this theory is why would T5 allegedly rig games of players who have filled the holes in their collection for the decks they care about? If you no longer have any incentive to buy more packs from a particular set of packs, until the new expansion, having your games rigged isn't going to suddenly make you want to do something (ie more pack buying) that you don't care about.

    Look at this: https://imgur.com/a/r3GAoSg

    Look at the global statistics up top.  They show you how other players have done against each class using the EXACT deck I'm using (it's actually quite popular so there is a ton of games by thousands of players in those stats).   So, now lets take a look at the pie chart below which shows what classes I have been paired with.  The most frequently faced class by far is Demon Hunter which has the lowest win-rate, by far (43%).  Out of all 10 classes I faced DH 29% of all my games.  Followed by Mage and Hunter which happen to be the next two lowest win-rate match-ups for that deck, in that order.  On my second most played deck that I linked at some point earlier (a Tempo DH deck) my most frequently faced class is Hunter which is also the lowest win-rate by far at (42%).  

    Consider then a player is having a bad night and feels like they are getting absolutely smashed by a "weirdly high amount of" (insert toughtest-to beat class for their deck here).   They feel compelled to make their own deck of that kind or one to counter it.  They que up and now get a "weirdly high amount of" (insert toughtest-to beat class for this NEW deck here).  They may feel compelled to make their own or counter it... but then what if... you catch my drift. 

    But I honestly don't think its just about buying more packs.  If there is a super broken deck that is winning +50% of games against all but one class (like that deck above actually) making some of these changes would serve to improve "player experience".  It would make matchmaking less flooded by the current meta and a good deck will still be giving a decent player 60% or more win-rate.  That alone seemed like a good enough reason to me for them to do it. 

     

    Again, I'm not trying to insult anyone.  I'm just being as honest as possible since I used to feel similar to many of you but these stats are crazy to me lol.  And I think its naive to believe there is no incentive for it to be done this way.   Now whether you believe whether or not "they would ever do that" or if the risk is "worth it to them" is a different story though for sure.  But those are all opinions and everyone will have their own imo.  

     

    Seriously would love to see some more match summaries to see if they look like mine.

     

    Thank you for your replies.  

    Much love,  jaypee

     

    EDIT: changed because it said nobody said it was rigged, but OP did say rigged lol. my bad.  

    0
  • Claeshj's Avatar
    90 13 Posts Joined 01/10/2020
    Posted 3 years, 8 months ago

    I have been farming a lot of DH games for portrait, and I am not seeing the same patterns as you Jaypee.

    All 3 decks have been mostly played at top1000 legend, 2 in standard but the Odd deck in wild. Data sets are: 158+239+109 games.

    Standard fast tempo DH:


    https://imgur.com/a/cdvWtTI

    Worst matchups: Shaman <1%, warrior 15%, priest 11%

    Most played: Demon hunter 25%


    Standard slower tempo DH:


    https://imgur.com/a/3ovthCF

    Worst matchups: Druid 10%, hunter 13%
    Most played: Demon Hunter 18%, rogue 18%

    Wild Odd DH:


    https://imgur.com/a/x9CFEJ3

    Worst matchups: warrior 7%, warlock 9%

    Most played: Mage 23%, Demon hunter 18%

     

    Edit: Added worst matchups and most faced opponent for easy overvie

    1
  • Almaniarra's Avatar
    HearthStationeer 950 1468 Posts Joined 03/21/2019
    Posted 3 years, 8 months ago

    This is not psychological neither rigging. This is one of the way which the game devs tries to balance the game with matchmaking. There is a simple fact for card games. Decks' visible winrates must be under %60 ( or should be at %50 ). If not, they need to nerf or buff some cards which they don't want to do most of the time because of market stance, balancing issues etc.

    It is also seem rigged vice versa. I mean, you can see that sometimes you only face the decks that you can counter but no one whines about that aren't they?

    Even MTG devs confirmed that they are using some rigged system for it.

    If your winrate with a deck is lower than the crowd, you mostly encounter with the decks you can counter or vice versa.

    So, Even Hearthstone devs don't confirm this, A card game must (and should in my opinion) use a system like this and Hearthstone has it too.

    Why did you think really they are keeping your game stats ? Especially for this. But this is agreed for other players too, not just for you. So this is not rigged game. It is equal for everyone. It applies you and me and others. All of the players who plays this game.

    You guys should accept this as a balance, not for rigged game. This is how you balance things. Matchmaking system is not only about matching same ranks. It also tries to match the decks.

    Those screenshots on OP proves that your winrate with that deck is higher than expected in the rank you are in so matchmaking system tries to balance it.

    Community should be accept this in my opinion or shouldn't be two-faced about how they win in a row against the decks that they are countering or should whine about those too.

    That's why some people suggest that "Always make some changes in netdecks while using them, you'll see that your winrates will be higher if you do this." because those changes are unexpected by this system and This is why some homebrew decks are successful at its first days and their winrates decreasing in time parallel to their popularity. This is why Rogue Decks(Not Rogue class) couldn't exist in this system. Sooner or later it would be discovered and because it is known from system and mirror matches, it is sentenced to lose its winrate.

    So, the system wants your winrate around %50 and you are trying to beat it. That's what you are playing. You are not trying to beat people who plays the game, you are trying to beat the system to gain your ranks. Strategy is that how you can beat this. This is not personal. You can use netdecks and beat it with your pilotting skills, you can create a succesful homebrew deck to beat the system or you can tinker netdecks to trick the system, you can stick with one deck for 4 months and beat it with your patience etc. There are so much ways but it is between you and the system. It is not like rigged for you but not rigged for anyone else. It is agreed for everyone and we all are trying to beat it. So everyone is equal. This is rigging at first glance but equality in it makes it "balanced". All players are equal and all are random, you should think.

    TL;DR: It is not a rigged game, It is a known balancing feature.

    Unpopular Opinion Incarnate

    0
  • CursedParrot's Avatar
    640 720 Posts Joined 05/29/2019
    Posted 3 years, 8 months ago

    I don’t see why HS would ever make matchmaking based on decks when they can just use MMR. If a player’s win rate is too high, they simply face better players and then they’re winrate goes down. Blizzard has no need to use some complicated system to decrease the winrate of decks. Also, if blizzard did have this system, how were there so many metas with high winrate decks? For example, Galakrond Shaman has extremely high winrates before it was nerfed. 

    On your point about decks decreasing in win rate over time, there are multiple reasons for this that don’t recruiter engineered matchmaking.

    1. As a high-winrate deck gets older, more bad players hear about it and start playing it, causing the overall winrate of the deck to decrease. For example, when Patron warrior was really good, many players who didn’t know how to play it well started playing it, reducing the overall winrate. 
    2. As a deck becomes more well-known, people learn how to play around it.

    3. When a deck becomes a large part of the meta, the meta adapts and more people play counter decks, reducing the winrate. 

    If your theory were true, there would likely be evidence for it beyond a few anecdotes.Also, HS has had a lot of nerfs recently, so I don’t think that they’d implement this to promote meta health.

    -1
  • jaypee's Avatar
    Uther 70 5 Posts Joined 07/26/2020
    Posted 3 years, 8 months ago

    This makes a whole lot of sense, and I agree with you that it does not necessarily mean its "rigged".   After all, good players would still have positive win-rates and they would still climb the ranks, just much more slowly.  Bad players will still have negative win-rates and would fall the ranks much more slowly.   Notice how you cant even fall below a certain rank once achieved... this can't be a "player experience" based decision right?  That's not a thing apparently.  What about giving lower win-rate players a bias toward nicer match-ups so they don't lose SO much?  Maybe then they have a better experience and keep playing.  Sure they wont be high-rank and still likely have a negative win-rate, but they wont lose so much.  Anyone who does not see how this sort of a system would massively improve the experience for more casual players (that are likely the bulk of HS players/clients) needs to think about this for a few more minutes.   

     

    For example in my Galakrond Warlock deck I linked earlier, where 29% of my games have been versus Demon Hunters:

    Once I learned the deck, it seemed like I couldn't lose playing in Silver/Gold/Plat (where I live basically LOL).  The win-rate got to 65%-70% about 20 games after I "learned the deck".  Keep in mind that was after losing many games "learning the deck".  So my win-rate for those ~20 games that brought me to 65-70% was quite high.  I suddenly felt that I started facing demon hunters NON-STOP at some point.  Salty as hell one night, I checked the match history and it simply did not confirm my suspicions.  Historically I was actually barely facing Demon Hunters more than any other class (had ~40 games total on the deck at this point).  Almost 100 games into the deck now, with the majority played while maintaining a 60%+ win-rate, DH opponents account for almost 1/3 of my lifetime games for that deck (which is the worst match-up for that specific deck w/ a ~43% win-rate).  I was mind-blown.  The top 7 best match-ups against my deck only account for 37% of all the opponents I have faced.

    I look at the classes I faced using other decks, like my DH tempo deck, the distribution looks nothing like the my Galakrond Warlock deck.  But I'm playing all of these decks at the same ranks, and often I literally alternate between them throughout a HS session.  So its not that there is simply a ton of X or Y class queuing at the same time as me... so its not a concentration of a specific class at my rank.... what is it then?

    What's with the super skewed distribution?  Why do my lower win-rate decks have a MUCH MORE even distribution of games spread out between opponents in all classes?

     

    Decks: 

    Galakrond WL

    https://imgur.com/a/r3GAoSg

    Tempo BH

    https://imgur.com/a/uPQShMY

     

     

    Also shout out to the user who posted their deck data as well @Claeshj .  It showed a difference from what my matchmaking data showed.  Others' summaries are exactly what I was hoping to see.  I just don't know what program they were using to track and therefore I don't know certain important facts like if all the games were played with the EXACT same deck to the tee, and what the global win-rate statistics are for that deck against each class (not just the posters' win-rate).  He is also Top 1000 legend which for many reasons is way different than the plebeian Gold/Plat ranks for the masses that I live at.  I would love to see similarly ranked players' stats too if possible (shoutout my fellow noobies) too.Particularly anyone using a deck that is getting them 60%+ win-rates or substantially higher win-rates than the global averages for that deck.  PLSS.  

     

    Thanks and much love,

    jaypee

    0
  • Almaniarra's Avatar
    HearthStationeer 950 1468 Posts Joined 03/21/2019
    Posted 3 years, 8 months ago
    Quote From CursedParrot

    I don’t see why HS would ever make matchmaking based on decks when they can just use MMR. If a player’s win rate is too high, they simply face better players and then they’re winrate goes down. Blizzard has no need to use some complicated system to decrease the winrate of decks. Also, if blizzard did have this system, how were there so many metas with high winrate decks? For example, Galakrond Shaman has extremely high winrates before it was nerfed.

    While players' winrates are important, decks' winrates are important too because it will cause to compress meta to 1 or 2 decks and Game would be unplayable. I mean, fun-wise. It will be like Play Jade Druid or Pirate Warrior if you want to win, nothing more or less. So it will decrease the fun feature of the game and it will be no more a game but just will be a competition because Hearthstone rewards players only if they can win so community of this game inclined to play for win, not for fun. Just trying to balance the game only around player winrates is not a solution for a healthy game because people inclined to abuse it always. Like you say, There were so many metas with high winrates and all of them saw nerfs at the end or got new counter cards on newer expansions just like Skulking Geist for example. That's why Teams are changed and people left from developing at the first place in my opinion.

    Quote From Author

    On your point about decks decreasing in win rate over time, there are multiple reasons for this that don’t recruiter engineered matchmaking.

    1. As a high-winrate deck gets older, more bad players hear about it and start playing it, causing the overall winrate of the deck to decrease. For example, when Patron warrior was really good, many players who didn’t know how to play it well started playing it, reducing the overall winrate. 
    2. As a deck becomes more well-known, people learn how to play around it.

    3. When a deck becomes a large part of the meta, the meta adapts and more people play counter decks, reducing the winrate. 

    All of these articles are true and I agree with them and I want to add 4th to these.

    4. When a deck gets higher winrates among others, it becomes popular and people tends to play these decks so it increases the Mirror Match posibilities and in mirror matches, There's only %50 chance. A deck always has %50 chance to win against itself. Even if its winrate is %80, its popularity will try to make its winrate near %50 because they will match with each other.

    Quote From Author

    If your theory were true, there would likely be evidence for it beyond a few anecdotes.Also, HS has had a lot of nerfs recently, so I don’t think that they’d implement this to promote meta health.



    I don't believe in coincidences. I know little samples are not evidence in statistics but i'm playing this game for 6 years and I'm experiencing it for 6 years aswell. It's not only for this game, I've experienced it in other card games aswell and I read around web about it so much times. If there are so much players that experienced this feature, it can't be just a coincidence. Also, Even some game developers accepted that they use a system like this in some conversations and interviews which are not from Blizzard or Hearthstone. Those conversations and interviews means that a system like this exists. We just can't be sure if it is applied to Hearthstone or not because Hearthstone devs don't say anything about it but these random encounters around web (Like OP) with my own experiences says that to me it might be.

    Btw, don't get me wrong, I'm not whining about it. I even think about it as a healthy feature.

    And lastly, You may say that These all are nonsense just because of the fact that Feature like that would be meaningless because of one of my reasoning which is "Winrates of decks would be near %50 in time sooner or later according to their popularity." So you can say that only player winrates matter. That's also true, but not in a healthy way. As I mentioned at my first paragraph, it will cause an unhealthy meta which will force people to play one deck or another. So best way to balance the subject between competition and the game is balancing all of its features which are Winrates of Players, Winrates of Decks, Card Balancing, System Balancing. I can't prove my points because i can't see all of the records but trying to balance only player winrates doesn't sound wisest and healthiest thing to me.

    There is also some hints and keywords in developer interviews and patch notes that might be evidence of a system like this, like for example "We saw that X card acts more efficient than we expected in Highlander decks so we decreased its Health and Attack value to make it more balanced." or "We know, playing against decks that mess with opponent's deck sometimes feels frustrating but We didn't change X cards for now but we're watching its winrates." etc. These all mean that they are not trying to balance only Player winrates. That would be even meaningless because A player's winrate might be %80 and it might be normal. The thing that would not be normal is that a deck's general winrate passing %60.

    I'm not writing this with my player side, just trying to approach it obectively as trying to think like a developer.

    Unpopular Opinion Incarnate

    0
  • Leave a Comment

    You must be signed in to leave a comment. Sign in here.

    ODYN
    0 Users Here