Understanding Match-up Theory, Part Two: 'Recognition' & 'Adjustment'

Introduction

Reading Tells I: What are they?

‘Reading tells’ is very straight-forward business (most of the time): All we are doing studying our opponent and are searching for any titbit of information that may reveal what deck they opponent are running. And the faster we do that, the better.

In the spoiler below, I have provided a quick proof to show that speed matters. If you don’t need any convincing that speed matters feel free to skip it.

[spoiler title =”Proof that Speed Matters”]

When playing against ‘Handlock’, on Turn Two the Warlock will usually decide to Life Tap (whereas a’Zoo deck’ would typically play a minion). Those two behaviours (life tap, playing a minion) are ‘tells’ and coming to conclusions about what they may signify (life tap = Handlock, Minion = Zoo) is what I call ‘reading’.

An ‘Average Player’ might therefore discover that they are playing against Handlock on Turn Two (due to the aforementioned behavioural tell). But a ‘Strong Player’ might have discovered what the match-up is before Turn Two! How? Well, ‘Handlock’ tends to Mulligan more aggressively than Zoo does (See here for more detail).

Let’s be a little hypothetical: Suppose that ‘PLAYER ONE’ and ‘PLAYER TWO’ both play 100 games versus Warlock with Deck X. In every single game PLAYER ONE figures out it is ‘Handlock’ during the Mulligan Phase, meanwhile PLAYER TWO figures out it’s Handlock on Turn Two.

Now here’s the key question: “Assuming all other things being equal, which of these two players is likely to report the better win-rate?”

Well, I think Player One is probably winning more games in this match-up, the simple reason being is that he/she can keep/throw cards based upon the assumption that they are playing Handlock (meanwhile, Player Two is searching for/keeping good anti-zoo cards). Therefore, on average, Player One will start the game with a better hand than Player Two. Clearly then, this hypothetical example demonstrates that the speed at which you figure out what you are playing against matters.[/spoiler]

Question: “How quickly can I find out what deck my opponent is running?”

The (generalised) Answer: To figure out the answer all you have to do is two things:

Create a list of all the decks in the opponents class that they could be playing.
Think about all the differences between those decks (different strategies, minions, etc) and try to identify at what point in the game where those decks are likely to deviate from each other. Sometimes decks may deviate at the mulligan stage, but at other times the deviation could be much, much later.

I have already given you an example of this above; Zoo plays minions on Turn Two, Handlock rarely does that. A second example would be to observe that ‘Grim Patron Combo Warrior’ will often play Fiery War Axe on Turn Two, even when there are no minions worth attacking. A Control Warrior meanwhile, typically prefers to use Hero Power, in order to buff Shield Slam. With that said, Grim Patron Warrior may even skip Turn Two altogether! (I elude to why here and here), whereas a Control Warrior is highly unlikely to pass up the armour Hero Power generates.

There are countless ‘tells’ to discover. However, since this series is aimed at providing tools and not answers to meta related questions I shall not elaborate any further. Consider it homework. 🙂

Okay, let’s move on…

Reading Tell’s II: What do they think we are?

Reading Tells I Predominately focused on how (and how quickly) we can decipher what deck our opponent is running. But in this section we are going to tackle the same question but from a different perspective; “Does our opponents play reveal what they think our deck is?” And if so; “Is that knowledge exploitable?”.

I shall go into depth about how we may go about exploiting our opponents assumption(s) for fun & profit in the next section. But for now, let’s consider what sort of plays may reveal what match-up our opponent thinks he/she is playing:

[spoiler]The BGH play reveals that the opponent expects you to be an Aggro deck that does not run anything with 7 or more attack (in the next section, I go into more detail).

[/spoiler]

[spoiler title =”No Mage/Druid Hero Power on Turn Two”]

What does the ‘no Hero Power’ reveal?

Suppose you are playing as a Warrior and your Druid opponent does the curious thing of not dealing one damage to your face. What could this mean? Well, it could mean that they have read my article and have decided to play-around Battle Rage. And since Battle Rage is only played in Grim Patron Warrior this means that they believe either that you are more likely to be a Patron Warrior than you are to be Control Warrior and/or you believe that playing-around Battle Rage is so ‘low cost’ you might as well do it anyway (regardless of how unlikely it is that you are running a Grim Patron deck). For example, since Freeze Mage does its first 15 damage via Alexstrasza deciding to not do one point of Damage with a Turn Two Hero Power has almost no downside for the deck.

Furthermore, Control Warrior is a terrible match-up for Freeze Mage but Grim Patron Warrior is actually beatable. Therefore, in my opinion ‘Freeze Mage’ should always make the ‘Anti-Grim Patron play’ even if the Freeze Mage player is 90% certain they are playing against Control Warrior.

I would however like to offer a word of caution; be careful not to misidentify tells. Notice for example, that I mention Mage and Druid. Not Rogue. Why? Well, the answer is very simple: although it could be the case the Rogue Daggers up on Turn Two and then passes without attacking because they wish to play-around a potential Battle Rage from a Grim Patron Warrior this play is, I’m afraid, more likely to signify that they have Deadly Poison in hand.

[/spoiler]

So I have given you two specific examples of plays that an opponent may make against you that reveal what they think you are playing. I could of course give you more examples but once I again I shall refrain from doing so on account of the fact that this article is intended to be theoretical and not meta specific.

Your list of homework activities is piling up, isn’t is?

Anyway;“How do we identity what a ‘match-up tell’ looks like (in general)?”

Create a list of all the decks in your class that you could play.
Think about the cards in your opponents deck
Identify card(s) (if any) that have different roles in the various match-up’s.
Wait for them to play one of the cards mentioned in (3): In what role do they deploy it? The role reveals your opponents match-up expectation.

I realise that this is all a bit vague so let me give a very simple example: An Ironbeak Owl being played on Turn Two suggests that the opponent is not expecting to get much value out of the silence effect, and would just rather use it as a 2/1 minion to trade. If you are a Paladin, this play therefore suggests that the opponent expects you to be some sort of Paladin aggro deck. And indeed, if you look at Joseph’s list there isn’t really much value to be had from a Silence (except for countering Blessing of Kings and maybe Coghammer‘s Battlecry). Had your opponent expected you to be a Control deck it is highly unlikely that they would have played the Owl on Two. And this is basically because every Paladin Control deck in the history of Hearthstone has run Tirion Fordring; and silence is the best counter to him.

To reiterate a previous warning I would like to remind you about taking care to not misinterpret tells: Your opponent may indeed be expecting you to be a Control Deck, they just dropped an Owl because they had two in hand and didn’t think that they would need both.

Okay, so these first two sections have mostly been about how to spot the ‘tells’ that reveal information about the Match-up. The next section aims to try demonstrate why we should make it difficult for the opponent to ‘read us’.

Stalling & Practising Subterfuge

‘Stalling’ & ‘Subterfuge’ is all about trying to delay giving the opponent information and/or trying to mislead your opponent about whatever you can:

Confuse them into thinking you are playing ‘Control’ when you are are in fact playing ‘Mid-range’.
Keep them perpetually confused about what cards your deck contains.
Keep them perpetually confused about what your hand is and what your game-plan is.
…And so on…

By misleading them, the hope is that your opponent will use key cards in the match-up inefficiently and/or make strategical mistakes.

Imagine for example that you are running Face Hunter versus a Druid. The Druid may, on Turn Three opt to play Big Game Hunter. In most situations this is an okay play versus Face Hunter, it is okay because you rarely get value out of the Battlecry in this match-up. But let’s now assume that you have ‘tweaked’ your Face Hunter list and have included a few late game cards, such as Dr. Boom in it. Had your opponent known that you tweaked your Hunter deck in such a way there is a very good chance that the Druid may have made some other play on Turn Three. But crucially, they didn’t know! And hence they used a key card of theirs inefficiently and may suffer later on as a result. This then, is a fairly simple example of a Hunter exploiting the Druid’s assumptions about the meta/match-up for fun and profit.

THE DRUID: “Ah, this is a face Hunter…” [Recognition phase] “…This sort of deck doesn’t run 7+ attack minions. Therefore I might as well play BGH now” [Match-up Adjustment phase]

(Hunter plays Dr. Boom)

THE HUNTER: “U Mad Bro?”

So this was an example of trickery at the meta-level; it turns out that due to the Hunter’s little bit of ‘meta-gaming‘ the Druid failed to adjust to the match-up properly. This is, incidentally, one of the most powerful arguments one can make for building your own decks; against ‘Home-brew creations’ (or against well established decks that you have personally ‘tweaked’) the propensity for opponents to ‘mis-step’ is significantly higher (when compared with ‘net-decked’ lists).

So that introductory example aimed to illustrate the idea that misleading your opponent about the nature of your deck can yield tangible and positive results. But this was done by tweaking a deck. In the next few examples I will showcase how you can gain an edge versus opponents with things you can do within a game. I’ll put these ones in spoilers.[spoiler title =”Delaying Tells”]

First things first. Let me link you to three Mage lists:

Okay, so as of writing this guide these are three of the most popular Mage decks you will see on Ladder and in Tournaments. For the purposes of this article there is only one thing I want you to pay attention to: All of these decks contain Mad Scientist. Okay, now let’s check out a possible position a Mage could find themselves in.

YOUR HAND: {3 cards} Mad Scientist, Mechwarper, Clockwork Gnome

OTHER NOTES: You have 2 Mana. Assume that both players skipped Turn One.

This is not the In-Depth Play Analysis series, and so therefore I am not going to go into crazy detail about what the “best play” is. The only thing I want you to understand is that we are a Mech Mage with two plays this turn:

Drop Mad Scientist
Play Mechwarper + Clockwork Gnome.

For the purposes of this article I only wish to bring one Salient detail into focus; namely that play #1 ‘delays the tells’ whereas play #2 clearly screams at our opponent that we are a Mech Mage.

If we make play #1 then our opponent simply does not know what deck we are playing and this may give us some sort of advantage. For example, by playing Mad Scientist a Druid is disinclined to Innervate-out Loatheb. Why? Because Loatheb is a critical card versus Freeze Mage (as I have already mentioned); to innervate it out on Turn Three is to use the card sub-optimally in that Match-up. Conversely however, Innervating Loatheb is a rather strong play versus both Mech Mage and Tempo Mage, owing to the simple fact that both of those decks like to get early board control (and Loatheb, merely as a 5/5 minion, denies the Mage that control).

This then, is a fairly simple example of how the lack of unambiguous information has posed a difficulty to the Druid; they must make a safe play, take a risk (such as play Loatheb and just hope we are not Freeze Mage) or ‘fish’ for more information (unfortunately for the Druid however, this ‘fishing’ takes time and precludes then from making a strong tempo play). In either case then, denying our opponent knowledge about the match-up can only benefit us.

The critical question however is how much the denial of knowledge benefits us — And make no mistake, this is an incredibly hard question to answer. The problem we have is that while the Mad Scientist play denies information (which is good) it is also a much weaker ‘tempo play’ (when compared with play #2). This then begs the question: “How do we win X match-up with Mech Mage?”, “In X match-up can we afford to play a bit slow or must we flood the board as fast as possible?”

In short: “Do we harm our ‘win-condition’ by playing Mad Scientist?”

I do not profess to know the answer. And in any case we do not need to know it either; the key point to grasp here is that there is a potential tension or ‘trade-off’ between denying information and making plays that reveal information but are otherwise strong.

For example, A Face Hunter reveals that they are a Face Hunter deck by playing Leper Gnome on Turn One. But this is a deck where every point of damage counts, therefore the decision to keep the Leper Gnome in hand (in order to deny the opponent match-up information) probably losses more games than it wins.

Okay how about one more example.

Both Control Warrior and Grim Patron Warrior use Acolyte of Pain, but Frothing Berserker is used only by Patron Warrior. Ergo, if you are playing Grim Patron Warrior and on Turn Three you have a choice between dropping the Acolyte or the Berserker you should give some weight to the notion that Berserker reveals considerably more information about your deck than playing Acolyte of Pain does. Ergo, the ‘denying information play’ is to play the Acolyte.

Maybe in this case you could also justify playing the Acolyte first for other reasons as well. For example, playing the Acolyte first may ‘bait’ removal that would have otherwise been spent on the Berserker. Moreover, if the Acolyte survives and there is a minion to attack then the Turn Four Beserker can hit the board as a 4/4 instead of being the 2/4 is would have otherwise been had you played Berserker on Three and followed up with a the Turn Four Acolyte. So maybe, just maybe, this is a case where we can justify the ‘denial of information’ on the grounds that Turn Three Acolyte is a good play in its own right (i.e. Protecting/Buffing the Berserker is a good thing). In this paragraph I’m trying to explain a rather simple (but important) idea: ‘denying information’ is merely one factor (among many) that you should consider when making a play.

[/spoiler][spoiler title=”The Warrior Ruse”]

In Th3 Rat’s Grim Patron Warrior Guide he advises that on Turn Two you should just equip Fiery War Axe, even if there are no targets to attack. Meanwhile, Control Warrior (when fighting against other control decks) typically prefers to use Hero Power rather than dropping the Axe in such situations, preferring to buff the Shield Slam.

Okay, so let’s suppose we are playing Control Warrior and we have the choice on Turn Two to Armor Up! or develop the Axe. Now here’s the interesting question:

“Is there any advantage for us to pretend to be a Grim Patron deck by equipping the Axe?”

For the sake of argument, let’s suppose we are playing at a high enough rank that this ‘tell’ is going to be understood by our opponent.

Again, just as with the Mech Mage example above it is hard to know whether tricking the opponent into thinking we are a Grim Patron deck is worth making a play that runs counter to the general ‘High Armour = powerful Shield Slam’ game-plan of Control Warrior. But with that said, can I construct an example that would demonstrate a possible advantage we could obtain via this act of subterfuge? Why yes, I can.

Suppose we are playing against a Druid and we play Acolyte of Pain on Turn Three (this play– unlike a card like Shield Block— is consistent with our previous ‘We are not Control but Grim Patron Warrior’ ploy). Let us now suppose that on Turn Four the Druid has a choice between Wrath and Swipe. Against Control Warrior the Druid would probably use the Wrath to kill the Acolyte; it’s more efficient and keeps the more powerful removal card in hard (after all, why use a bazooka to kill a fly?). But what about against Grim Patron Warrior? Well, in this case they may remove the Acolyte with the Swipe! Ergo, by tricking our opponent into thinking we are Grim Patron we have convinced them to use Swipe inefficiently; does this advantage offset our lack of amour we could have otherwise obtained via Turn Two Hero Power? Maybe.

I suspect that some of you would like a little bit of an explanation as to why the Druid may Swipe here. Okay, I’ll oblige (albeit grudgingly). The problem with Swipe in the Grim Patron Match-up is that the one-damage to all enemies can often be more harmful than beneficial; Swipe just gives the Warrior more Grim Patrons, more cards (Acolyte of Pain, Battle Rage), and possibly more damage (e.g If you Swipe something other than the Frothing Berserker). Another subtle point to note is Grim Patron Warrior often likes to equip (and attack with) Death’s Bite on Turn Four in order to set up a strong Grim Patron Turn Five (attacking with the Axe again yeilds a 3/3 and 3/2 due to the deathrattle). This play is often particularly potent versus Druid owing to the Druid’s lack of board clear. Notice that if you Swipe the 3/3 the 3/2 becomes a 3/1 and spawns a 3/3, so clearly Swipe sucks here (in effect, Swipe merely did one damage). Whereas Wrath clears the 3/3 easily, and now if you have a Keeper of the Grove you can kill the 3/2 Patron as well. In short: the Druid finds a way to counter the Warrior’s potent Turn Five with a six mana two-card combo. But crucially, such a combo was only possible because the Druid preferred to use the Swipe on the Acolyte (as opposed to using the Wrath) a few turns before.

Once Again, this leaves us with an interesting but profoundly difficult question: When we played the War Axe as Control Warrior we were being slightly inefficient (We are failing to follow the general warrior ‘get amour’ game-plan). But this ruse also lulled the Druid into too using a card inefficiently. So here we have dynamic trade-off (both players have been inefficient, in some sense) and the question is who stands better after the dust has settled. Against Druid, maybe the ‘I’m not a Control Warrior’ ruse has merit, but maybe against Handlock you are perhaps simply better off getting the Armour since that helps you deal with the Warlocks powerful Turn Four/Turn Five plays (e.g. Twilight Drake, Mountain Giant) moreso than the Axe does. And in anycase, so what if Warlock thinks your are Grim Patron Warrior; are they going to play the first 5-6 turns of the game any differently? Probably not.

But anyway, since this Match-up series is supposed to theoretical I really don’t want to spend any more time explaining highly advanced techniques specific to particular match-ups (chances are that most of the above will be out-dated information within a month, anyway), hence I only offer these explanations grudgingly. 🙂

Before leaving this example it is perhaps worth pointing out the example could also work the other way round: If we were a Grim Patron Warrior there may in fact be some merit to trying to convincing the opponent we are a Control Warrior (we do that by playing Armour Up! on Turn Two instead of developing the Axe). Thus, in this case when we play Turn Three Acoylte of Pain the Druid maybe more willing to burn the more valuable card in the match-up (i.e. Wrath) due to him/her believing that we are more likely to be a control deck; and now, with the useless swipe in hand, the may Druid struggle to combat the ‘Grim Patron + Enrage’ Turn Five play.

All this stuff is profoundly complex and hard to evaluate. Therefore, I simply wish to end this discussion by making a very general (and ‘safe’) claim: In some match-ups, there is some merit to delaying tells and/or even pretending to be another deck altogether. And as general rule of thumb this is more effective in polarised match-ups (e.g, If Deck Y plays against Decks A and B in very similar way, there is little advantage to be had by masquerading as deck B when you are in fact deck A. Conversely though, if Deck X plays against decks C and D in a radically different way then the is potentially a large advantage to be gained by pretending to be Deck C when you are in fact Deck D). [/spoiler]

Even Pro’s make these mistakes (sometimes)

Let’s set the scene….

ESL Legendary (Grand Finals)

GAME 1: PinPingHo (Shaman) vs Zalae (Warlock).

In this game PinPingHo misread the match-up and paid the ultimate price. One can only assume that thought that he was playing against a Handlock build (which is why he was so reluctant to use Hex; he was saving them in order to counter the Mountain/Molten Giants that are not actually in the deck! Had he of known, it is likely that he would have hexed a Sludge Belcher). In short, PinPingHo fundamentally misunderstood both the Warlocks ‘Win Condition’ (which perhaps explains why he didn’t Hex one of his own minions on the penultimate turn of the game) and his own role the match-up.

Here’s the game:

Conclusion

Okay, so that was a lot to take in, kudos for finishing it. I realise that many of you probably found this to be a difficult and dense read, and for that I apologise. Nonetheless, I hope that some of you have taken some idea’s with you that may help improve your match-ups. I hope to have demonstrated that identifying the Match-up as quickly as possible yields advantages (particularly if you recognise the match-up before your opponent does), and that, since such knowledge is useful it makes sense to try and deprive your opponent of it!

Anyway, I’m going love you and leave you now. As always, comments and likes (and Tips 🙂 ) are appreciated.

Understanding Match-up Theory, Part Two: ‘Recognition’ & ‘Adjustment’