MMR- Matching by skill or by winrate?

Submitted 3 years, 11 months ago by

I just climbed with a murloc deck from Platin 9 to Diamond 10 and encountered several Demon Hunters who just played badly. About two-thirds of them didn't read the card texts, and cleared Murloc Warleaders or Grimscale Oracles last, after taking unnecessary amounts of damage and losing minions to the extra +2 or +1 attack. One even played [Hearthstone Card (Warglaives of Azzinoth) Not Found] and cleared 4 murlocs, the Murloc Warleader last!

This made me wondering, how the Match making is done. Is the upper end of Platin still very low skilllevel, or is skilllevel just an euphemism for winrate? So instead of some weigthening algorithm, that evaluates how the player compares against decks of the same strength, MMR is just done by deck strength alone? Does anyone know more about the new MMR, and how strong decks play a role in this?

  • Elfensilver's Avatar
    595 663 Posts Joined 03/14/2019
    Posted 3 years, 11 months ago

    I just climbed with a murloc deck from Platin 9 to Diamond 10 and encountered several Demon Hunters who just played badly. About two-thirds of them didn't read the card texts, and cleared Murloc Warleaders or Grimscale Oracles last, after taking unnecessary amounts of damage and losing minions to the extra +2 or +1 attack. One even played [Hearthstone Card (Warglaives of Azzinoth) Not Found] and cleared 4 murlocs, the Murloc Warleader last!

    This made me wondering, how the Match making is done. Is the upper end of Platin still very low skilllevel, or is skilllevel just an euphemism for winrate? So instead of some weigthening algorithm, that evaluates how the player compares against decks of the same strength, MMR is just done by deck strength alone? Does anyone know more about the new MMR, and how strong decks play a role in this?

    1
  • Lemushki's Avatar
    Squirtle 1110 1025 Posts Joined 03/22/2019
    Posted 3 years, 11 months ago

    It is mostly based on winrate alone. If the deck is too good that you can pilot it with no thought, then the skill level gets lower.

    Lemushki - The one and only since the 2006 rebranding.

    1
  • Erodos's Avatar
    Crossroads Historian 945 1019 Posts Joined 05/29/2019
    Posted 3 years, 11 months ago

    How would you measure skill objectively?

    7
  • Elfensilver's Avatar
    595 663 Posts Joined 03/14/2019
    Posted 3 years, 11 months ago

    By comparing your winrate against the winrate overall of the deck you're piloting. For at least all the people that play meta decks, there is a lot of data to compare against. And if you really play a deck that is different to all metadecks by like 4 or more cards (just spitballing here) then by winrate alone.

    1
  • RavenSunHS's Avatar
    Refreshment Vendor 880 1487 Posts Joined 03/27/2019
    Posted 3 years, 11 months ago
    Quote From Erodos

    How would you measure skill objectively?

    This.

    Measuring skill is bound to too many variables, as in, how do you define a list within an archetype? Error margin would be too high.

    More likely, the people you encountered had a good MMR simply because they were playing DH, and getting good winrates despite poor skills or low attention...

    2
  • Suchti0352's Avatar
    Hero of Warcraft 890 1030 Posts Joined 03/27/2019
    Posted 3 years, 11 months ago

    We simply don't know how the mmr gets calculated and that's a good thing. Otherwise some players would  exploit it to win more games.

    5
  • DelkoHS's Avatar
    Child of Galakrond 485 481 Posts Joined 05/28/2019
    Posted 3 years, 11 months ago

    You can't measure skill though.

    4
  • Pezman's Avatar
    Staff Writer 2235 2220 Posts Joined 06/03/2019
    Posted 3 years, 11 months ago

    As others have noted, there is no reliable way to measure skill. But I've noticed the same thing as you, specifically with Demon Hunters. Here's my theory: new players with very small collections can access far more cards for Demon Hunter than they can for other classes, and newer players are more likely to make those clumsy mistakes. I'm certainly not trying to be insulting, and I've obviously made my share of dumb moves, even after having played for years. It makes sense, however, that many inexperienced players have been climbing quickly with DH, because of how OP so many of the cards are, even after 2 rounds of nerfs.

    "Be excellent to each other." -Bill and Ted

    7
  • OldManDice's Avatar
    145 11 Posts Joined 09/14/2020
    Posted 3 years, 3 months ago

    I know I'm late to the discussion but here's my take based on my experience with MMR in other games where it's a bit more transparent and my Hearthstone play experience.

    I'm fairly certain that MMR is calculated using your WR vs the WR of you opponents over a number of games with a modifier for your ranking (both highest in previous season and current). I don't think it takes into account the type of deck you are playing at all. 

    This method of calculating MMR (WR vs opponents WR modified by ranking) ensures your MMR has a floor cap (because of the ranking modifier) and fairly accurately pits you against players of similar skill levels. However I generally find that in the lower rankings I am playing less skilled players which only propels me into the higher ranking with higher skilled opponents faster. All things considered I think its about as fair as a system can be in this type of game.

     

    0
  • meisterz39's Avatar
    925 1200 Posts Joined 06/03/2019
    Posted 3 years, 3 months ago
    Quote From Elfensilver

    By comparing your winrate against the winrate overall of the deck you're piloting. For at least all the people that play meta decks, there is a lot of data to compare against. And if you really play a deck that is different to all metadecks by like 4 or more cards (just spitballing here) then by winrate alone.

    Do you have any evidence you can cite for this claim? I've never seen anything from Blizzard that suggests this, nor have I seen anyone else online claim that it's true.

    0
  • Leave a Comment

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

    ODYN
    0 Users Here