The Midrange Shaman Mirror

How to maximize AOE for fun and profit.




shamanleftshamanrightAh, Midrange Shaman.  Have you seen any of them on the ladder recently?  I feel confident in saying that you probably have – mostly because the deck is about a quarter to a third of the ladder and has a 55% winrate despite being by far the most highly represented deck in the game.  It does everything well – you’ve got undercosted early game minions, undercosted midgame minions, undercosted late game minions, premium targeted removal, card advantage, and powerful area of effect removal.  What’s the best way to beat this monster?

  1. Be brave and play freeze mage, hoping to dodge warriors and also all the other bad matchups.
  2. Make a metagame call, learn a deck with a 50%ish mid shaman matchup (like secret/face hunter), and navigate your games with tight technical play.
  3. Learn how to crush the mirror match!

If you want to hit legend this season, option 3. is likely the easiest way to get there before the new cards are released – and there are definitely ways to play your games in a way that will give you a strong advantage when playing as and against Midrange Shaman.

Key question – How can we gain an edge playing the best deck against the best deck?

Having played lots (and lots and lots) of Mid Shaman mirrors, I came to the vS analysts with the initial assumption that maximizing AoE value is how to win the matchup.  The numbers came back and there’s quite a lot of interesting data here to suggest that – let’s take a look and start with the big clear.

But first, a brief note about data

Before we go any further, please note — we are not dealing in absolutes here and data as a whole asks very specific questions.  We will learn below what generally happens when certain cards are played on certain turns — but we do not know what would happen if the cards weren’t played, nor do we know the context of the games in specific, decklist variations, player skill differentials, etc etc etc.  What we will identify are trends – when a player does a specific thing on a specific turn, generally this is what happens.  Plus, it is important to note that the underlying assumption is that when a player does something, it is the best move they could make, according to their assessment. However, we emphasize that the numbers should not be viewed a as gospel!  Great players are able to identify when to take a chance, buck the trend, play in an unusual or generally “suboptimal” way and claim victory from defeat.

Lightning Storm

lightningstorm1

lightningstorm2

We asked the database “what was the winrate of playing Lightning Storm on turn X” and put it together in a table that also included total number of games in which Storm was played on turn X.  We combined all late game turns with less than 100 games (Beyond turn 15).  (Spoiler: We’ll be doing the same thing for Maelstrom Portal in a second.)

Wow, take a look at the difference here – the winrate line trends straight up.  The differences between the low end and high end win percentage is a whopping 33%!  The chance of winning the game doesn’t trend above 50% until turn 8, at which point the likelihood of taking the win jumps turn after turn after turn.  This makes sense, if you think about it – turn 8 is the turn after Thunder Bluff comes down and represents a large tempo investment that can be reset with a big board clear.  We ran some averages of before turn 8 and after turn 8 and came up with the following:

  1. The average win rate when Lightning Storm is played before turn 8 – 40.9%.
  2. The average win rate when Lightning Storm is played after turn 8 – 57.7%.
  3. The average win rate when Lightning Storm is played after turn 10 – 61.5%

The difference between playing Lightning Storm before turn 8 and playing it on or after turn 8 is 17% on average.  If you can manage to get by without playing Storm until turn 10, that jumps to a 21% difference.

What does this information tell us?  Well, the obvious answer is “hold your storms if you can” – playing them early to gain tempo or press an early advantage locks down your midgame with overload, rewards your opponent for holding AoE, and prevents you from punishing your opponent from going wide.  However, this also tells us that forcing our opponents into using an early storm is extremely beneficial to us.

As a reminder from our above caveat about the data… there will be some games you win because you have to play an early storm and it works out.  There will be times when holding storm will lose you the game!  Again, this is a an average result, not a guarantee.  There’s also an interesting bit of data (albeit with a small sample size) — starting on turn 15, the lategame win percentage starts trending downwards, though we still land north of 50%.

Midrange Shaman doesn’t mind early pressure because of its resilience to midgame pressure thanks to free 5/5 taunts and ability to clean up messy boards.  However, one of the ways Midrange Shaman can lose is if enough early pressure comes down to put the life total under constant danger. While the overload cards are powerful, they can prevent you from dealing with threats in successive if you are constrained in your ability to respond.  There are some openings in the mirror that embody this kind of aggression – typically involving Totem Golem, Flametongue Totem, and/or Feral Spirits – where huge chunks of life can be carved out by turn 5.  If you are able to force lightning storm with an early aggressive opening, it will likely reward you – as long as you have the tools to refuel/rebuild afterwards.

With that being said, once your opponent blows a storm, your game plan likely has to shift a little bit back to the usual value-conscious approach unless you have the tools to refill the board and finish things quickly (e.g. Bloodlust).  Do NOT use Lightning Storm as an aggressive weapon unless you absolutely have to – refilling the board when they’re overloaded is good, but using storm to clear out a couple of taunts and bring them to a low life total is not.  You’ll likely get cleared and then you’re back to square one.

Now, there are some cards that reward an aggressive push… enter stage right!




Maelstrom Portal

portal1

portal2

The methodology is the same, but the results are quite different — the differential between the extreme outcomes is about 19%, which is still a big deal, but the average outcomes are much closer.  Here’s how they look:

  1. The average win rate when Maelstrom Portal is played before turn 8 – 49.9%.
  2. The average win rate when Maelstrom Portal is played after turn 8 – 55.3%.
  3. The average win rate when Maelstrom Portal is played between turns 4 and 7 – 50.9%.
  4. The average win rate when Maelstrom Portal is played between turns 8 and 10 – 54.7%
  5. The average win rate when Maelstrom Portal is played after turn 10 – 56.7%

Looks like our averages here are much, much closer – only about 5% or so.  If we cut out turns 2 and 3, the averages are only 4% apart from the key midgame turns of 4 through 7 vs the key lategame turns of 8 through 10.  This card is much more flexible and there may very well be games where using portal as an aggressive tool is not a bad call, especially if it will bait a board clear from your opponent.

The biggest jump in portal winrates is actually from turn 3 to 4 with a 6% jump, after which the rate of the climb stabilizes until you see some really big late game jumps.  Similar to how there’s a Thunder Bluff defensive swing turn with Storm, there’s a Wrath of Air totem offensive swing turn with Portal that corresponds pretty strongly with the 4th turn of the game… and there’s no overload to punish you for playing it earlier.

Note – please take the late games with a grain of salt, as about 75% of all games sampled had a portal played between turns 4 and 9 – those really late game situations may very well be indicative of slow starts from one or both players where portal was able to be significantly amplified and acted as lightning storm #3.  (More on this later.)

A juicy midgame portal clear that can take out a few minions is likely worth your time, especially if it forces your opponent to storm on turn 6.  That ruins the possibility of a turn 7 Thunder Bluff turn and sets you up to retake the board and punish your opponent’s overload.  Don’t feel a need to hold it forever and get super greedy – if you are able to keep the board in a stable state by all means do so, but you don’t want to miss your opportunity for portal to be played and be relevant.

So what can we do with this information?

To start, we can maximize our advantage in deck construction.  You’ll see many lists floating around from conquest tournaments and high legend finishes that only have one lightning storm – do not fall into this trap if you are grinding ladder!  Unless you are lucky enough to dodge mid shaman (and zoo, for that matter), you need both storms and both portals.  In fact, the best lists for the ladder likely include at least one Kobold Geomancer – if you can consistently amplify portals to 3 damage without RNG being required, you will have a significant leg up on your opponents and will be able to hold your storms for more effective massive clears.

During gameplay, we can master how to leverage removal tools and learn discipline.  It is generally in your best interest to fight for board control but not fear your life total until you get to a life total where Spirit Claws can no longer be used to effectively trade – it depends on the game, but typically between 20 and 15 is where your life total has to start being preserved.  As usual, you’ll want to build board control and seek favorable trades.  Make sure that your board clears are gaining you actual card advantage and aren’t just eating up a couple totems and a Mana Tide.

We can also assume that our opponent has read this article or figured out how best to combat the mirror and master the practice of baiting removal.  Think of Lightning Storm like Brawl – it’s very similar in the context of how we want to use it and how we want it used against us.  If you can build a board that puts some reasonable aggressive pressure on your opponent — think of a Golem->hero power->Feral Spirits opening — you may be able to pull some AoE out of them ahead of schedule.  The magic number is 3 – that’s the amount of health you want your minions to have going into the midgame.  2 health AoE clears are easy to find, but 3 health ones are much harder to build.




Cool!  Who are you, anyway?

Who am I?  None of your bus– I mean, I’m RidiculousHat, a 9-time legend player, sometimes streamer, and amateur Hearthstone tournament/league player with a background in CCGs (particularly Magic: The Gathering).  You’re welcome to check out more of my blathering at twitter.com/ridiculoushat or watch me (usually on Tuesday and Thursday nights) at twitch.tv/ridiculoushat.  Thanks for reading!

This article was brought to you by:

ridiculoushat-writerEndofDaysTzachi

 

 

21 Comments

  1. Why is there no mention of Azure Drake + Storm given that it’s a common 8-mana swing play that would contribute to the increased winrate in the later game?

    In addition, the later in the game, the more likely the game is already decided regardless of Storm/Portal. Assuming the data contributors are above average compared to the total player base, the longer the game, the more decisions, the more edge we get, and hence the increased winrate when “card x” is played on later turns.

    Another problem is the game can be won and people are BMing with Storm/Portal and then kill the opponent and the Storm/Portal erroneously “contributed to the win.”

  2. Very nice article. Thanks a lot.

    I would read all articles that are like this one. And not only mirrors.
    – Make it a thing:
    The matchup of the week. Or something like that.

    Greetings to all.

    • YES! It would be such a good idea. I’m always looking to expand my match-up knowledge but it seems like there aren’t many sites that do that. Please consider doing this a consistent thing :).

  3. What I’d be more interested about is the effectiveness of tech cards, like Ooze/Harrison, Argus, Bloodlust etc.

    • I agree, I think it would be fascinating – the core of the deck is pretty solid, but I’m not sure about variance between:
      -2nd mana tide
      -Argent Horserider (1 or 2)
      -Weapon destruction
      -Bloodlust
      -Rag/Al’Akir
      -Fire Elementals
      -4 mana 7/7

      Etc.

  4. This is cool but I think you skimmed over an important part of this analysis, which I would love if the VS team could examine further. The issue is that this data is hiding already winning or losing matchups. Part of playing lightning storm on turn 6 is BEING ABLE TO HOLD lightning storm until then, and if you’re not forced to use it, you’re not losing early i.e. you have a better chance of winning.

    Off the top of my head some approaches to filter this data more would be:
    -Health differential when played
    -Minions played when played
    -Turns held before play (turn played- turn drawn)
    -Spellpower when played
    I don’t know if any of this data is available, but considering even one of these variables would definitely help.

    You also only use averages in your analysis, having the variance would be good to. In a similar vein, if more data were available, you could look at more sort of clustering effects, like high health AND late storm is high wr, low health and late is low, etc etc.

    Another idea is to simply look at conditional probabilities pertaining to health and turns -before factoring in lightning storm timing. It could be that Shaman winrate already increases linearly with turns, and that lightning storm is basically ineffectual. It could be that, if above a certain health by a certain turn, winrate is much higher. This is even more important than clustering or anything, and with the data available you already have some of the tools to perform this analysis.

    The last and most grievous error is that you have two of both cards! You can play one early AND one late! Where’s that consideration? If Lightning storm and Maelstrom portal are so important in this matchup, surely it’s worth investigating the effect of playing 1 vs 2 vs 3 vs 4, no? And what about double counting? If I win and I played on turn 2 and turn 8, does my wr contribute to both? This is even more obscuring the actual effect with already winning players.

    I’m sorry if my criticism is harsh, not trying to attack you at all. Just happy to see data being used and hungry for more. Keep up the good work and analysis, and I hope you are given more data and tools to analyze in the future. This is the kind of stuff I want to see.

    Tl;dr: This data might just be signalling that losing early is correlated with losing in general, needs vastly more fine tuning and conditioning.

    • This is great feedback and I agree with you 100% – but unfortunately there are limitations to the tools available to us. At this time, the data we can pull is when a card was played and the correlated winrate. Health total correlation, turn plays/draws, etc etc etc are all not able to be referenced at this time. There are lots of situations where the data presented above may not be applicable or may be influenced by other factors in the game – trust me, I’d love to say “if you’re below 15 before turn 8 in the shaman mirror, winrate of an earlier lightning storm increases” or something along those lines.

      The highly analytical side of Hearthstone is in its infancy and I can’t wait to see what happens as the data gathering tools become more advanced. As it is, we have to stick to making broader conclusions and looking at trends as opposed to zeroing in on specific circumstances.

  5. Very good article, but i’d like some follow up on the early game… some plays like coin+golem >> trogg, or coining a totem against mage if you have fire totem in hand, are tactics that can up your winrate drastically, imo.

    • These are great notes! Keep in mind if we were analyzing every mid shaman matchup and providing detailed data and notes on all cards and turns, the article would be longer than War and Peace – generally my approach is going to be to dial in to more specific swing turns and provide analysis where data can guide us to make more informed decisions. I don’t know if there’s as much value to saying “here’s how good Mana Wyrm is on turn 1” for example – because I hope there’s nobody throwing back Wyrm in the mulligan. Instead, I want to analyze situations that have more play to them so that the data really helps to illuminate unclear scenarios.

  6. You forgot the importance of Hex in your analysis. Hex Ragnaros or Thunder Bluff Valiant os huge. Later I play Hex more chances I’ve to win the game.

    Need a mid range shaman vs tempo mage analysis. Can’t win it must throw everything quick to stay alive and they keep kiling my minions, Flamestrike and Fireball etc my face so my health becomes critical fast ad I die with a Roaring Torch gg ez for mage.

    • So Hex can be relevant in the lategame, but it’s not really a swing card that has a lot of play in the matchup and I doubt it has a ton of specific winrate impact other than “were you able to Hex the Rag” – and hexing a thunder bluff is good, but not enough if you don’t already have board control or not as relevant if you’ve already claimed board control. AoE is really what the matchup hinges on from my experience.

      The tempo mage matchup requires you to turn into the aggressor sooner rather than later and start pushing your own face damage, encouraging them to use their burn *defensively* – if it’s not targeting your face, you’re in a good spot. It plays much like the hunter matchup – mid shaman can always retake the board but has no life total recovery mechanism and needs to establish itself as the beatdown early enough that tempo mage feels threatened.

    • So the cards you definitely want to see are Trogg and Golem along with Spirit Claws. If you have at least one of those, you can keep Spirits. If you have Claws, you can keep Thalnos or Geomancer. If you’re on the coin, keep Lightning Bolt. I generally don’t keep AoE, but if I have the claws+spell power hand, I’ll consider keeping Maelstrom Portal on the draw.

      The goal here is to have some kind of early action – you don’t want early reactive cards and you generally do want early proactive cards. The matchup is all about incremental value and board control – Trogg and Golem both force early answers from your opponent and put them at a disadvantage, theoretically baiting early AOE. Spirits does the same thing but loses to Claws or Golem if you don’t have a turn 1 or 2 play. Bolt on the draw is good for suppressing their early Trogg or an early Golem with claws or a spell power 2 drop.

      Hope this helps!

        • It might not be, but it’s going to garner its best advantage on turn 1 or at least before turn 5 and it’s worth mulliganing for.

          We’ll try to see if the information as requested could make an article, but keep in mind that the data we have access to is when a card was played and the percentage of times that player won the game. I expect a pretty steady line starting high on turn 1 and dropping – might not have enough meat for an entire analysis.

          Thanks for the suggestion!

          • Important thing is going first and second also.i think trogg on 1 going first should be worse the trogg on 1 going second

  7. Nice article with deep data analysis, nice to see someone presenting useful information based on facts! Keep it up

Comments are closed.