Card game design: how to objectively prove the fun of a given card, mechanic, or design (using Hearthstone as an example)
Possibly one of the most common issues in card games is the inevitable complaints of gamers when a new set of cards is released.
“X card is broken! It’s way too powerful!”
“I hate playing against X card!”
“Why do you even print X card?”
Sometimes these complaints are justified, and sometimes not, but one problem always exists: how do you prove that your opinion on a card design is correct? How do you stop the inevitable cry of “well I like this card, and all fun is subjective” “each to their own” and other such statements?
It’s actually rather simple. When one understands the basic design principles of most card games, it becomes very easy to objectively prove if a card is ‘fun’ or not. The first thing to note is that whether a card is fun or not has no correlation to its power level. Cards of high, average or low power levels can all be objectively bad for fun.
In order to properly explain how this works, we will examine firstly some cards and decks that use OTK (one turn combo) strategy in order to win the game, as they are the main issue to be dealt with. Here is a very simple example:
Consider this card for a moment. When would you include this card in a deck? Outside of its deathrattle, it is an extremely poor card, something which can be proven:
- It gives no defense against enemy minions (it doesn’t have Taunt)
- It is easily countered by cheap removal spells such as Shadow Word: Death, Execute, or Sap
- It has no immediate impact on the state of the board or game
- It is vastly out-valued by almost all other neutral minions (examples: Y’Shaarj, Rage Unbound, The Lich King, Boulderfist Ogre…)
The reason I have explained the above is to demonstrate a logical proposition: no player would ever put this minion in a deck unless they intended to activate its Deathrattle successfully.
Of course, that is obvious. It is relevant because once you put this card in your deck, and you intend to activate its deathrattle, all other cards in your deck are worthless in comparison. Why would you try to fill your deck with threats to your enemy when you can rush to activate this card? All other cards that do not advance your goal of playing Mecha’thun, will simply hinder your game plan. If you plan to kill your enemy in any other way, having Mecha’thun in your deck simply means you will have dead draws sometimes and thus it is useless.
Thus, the second proposition: all OTK decks contain the exact same three groups of cards. Obviously, this does not mean literally the same cards, but the same groupings, of which there are always three:
- Combo cards (examples: Mecha’thun, Archmage Antonidas + Sorcerer’s Apprentice, Seek Guidance, Uther of the Ebon Blade)
- Stalling cards (examples: Frost Nova, Doomsayer, Cloak of Shadows, Ice Block)
- Draw cards (examples: Arcane Intellect, Coldlight Oracle, Research Project, Loot Hoarder)
This is due to the aforementioned fact: any other card you might place into an OTK deck simply hinders the rest of your gameplan. This can be proven for every OTK deck that exists. Let us take a few example decklists to demonstrate this point.
It’s worth noting that some cards often fit into two groups (for example, many stalling cards often draw cards or thin the deck). Cards such as Mad Scientist, Bloodmage Thalnos, and summon-from-deck spells which thin the deck such as Call to Arms often fit this description.
It is worth mentioning here that Mill decks are in fact also OTK decks; they do not immediately appear to be, but they are simply decks whose ‘Combo’ cards are overlapped with its ‘Draw’ cards, as their only way to kill you is to force you to die from overwhelmingly high fatigue damage (to the extent that, even if you could recover your full health every turn, you will still die from drawing too many cards).
- Combo cards (Archmage Antonidas; Sorcerer’s Apprentice; Molten Reflection; Emperor Thaurissan; Simulacrum)
- Stalling cards (Frost Nova; Ice Block; Ice Barrier; Blizzard; Ray of Frost; Doomsayer; Mad Scientist)
- Draw cards (Mad Scientist; Novice Engineer; Arcane Intellect; Coldlight Oracle; Sphere of Sapience)
- Combo cards (Uther of the Ebon Blade; Auctionmaster Beardo; Banana Buffoon; Emperor Thaurissan; Clockwork Gnome)
- Stalling cards (Time Out; Humility; Righteous Protector; Call to Arms; Consecration; Shrink Ray; Flash of Light; Wild Pyromancer; Doomsayer; Bloodmage Thalnos; Equality; Truesilver Champion)
- Draw cards (Bloodmage Thalnos; Loot Hoarder; Flash of Light; Solemn Vigil; Call to Arms)
- Combo cards (Valeera the Hollow; Coldlight Oracle; Lab Recruiter; Gang Up; Preparation; Counterfeit Coin; Prize Vendor; Brann Bronzebeard; Togwaggle’s Scheme; Shadowstep; Secret Passage)
- Stalling cards (Valeera the Hollow; Evasion; Vanish; Cloak of Shadows; Glacial Shard; Armor Vendor; Mistress of Mixtures; Sap; Preparation)
- Draw cards (same as Combo cards for Mill decks)
The above demonstrates that OTK decks all contain the same elements and strategy. This all goes to the root of the problem: where are the cards that are supposed to react to the other player? There are none.
The purpose of OTK decks is to effectively play a one-player game; your only goal is to acquire your combo cards and play them, as once you can do that, your opponent cannot stop you and the game is over. As a result, the other player’s actions are generally irrelevant; games become a coin toss, in which the game is largely decided by which deck you are facing rather than your decisions. To illustrate this, here are two kinds of decks that have vastly different win rates against OTK archetypes.
This is an archetypical aggro deck. Against OTK decks this deck has very high winrates; this is because the speed at which this deck can deal damage is very high, and because the mana curve is very low, allowing this deck to deal its damage before the enemy is likely to draw their combo pieces. Because OTK decks only ever have three groups of cards, they rarely have sufficient healing to prevent the aggro deck dealing fatal damage (and stall cards generally only account for minions or weapons already in play for a turn, which isn’t too helpful against weapons and Charge minions).
This is an archetypical control deck. It contains very little in direct damage and aims to out-value its opponent in the late game; because it cannot directly kill the opponent or present a sufficient threat until very late in the game, the opponent will almost always assemble their combo first, no matter the skill of the opposing Priest player.
In both of these examples a clear pattern can be seen: the skill and decisions of the players is largely irrelevant. When facing an aggro deck, an OTK deck will often lose no matter what decisions it makes; the reverse is true for an OTK deck facing a control deck.
This leads to the final logical proposition: a card game’s fun can be based, entirely — as with most strategy games — on how much the skill of the players contributes to victory. There is a reason people don’t play games like blackjack outside of casinos; card-counting aside, luck is the only factor that determines success there (and as such, the chance to win money is what makes the game worth playing in that context). A card game’s fun can therefore be said to be objectively measurable by the amount skill contributes. That doesn’t mean luck cannot be part of the game, simply that the outcome of the luck should be influencable by skill.
Now we can see the heart of the problem: excessive deck polarity. Deck polarity, to put it simply, is the difference in win rates that a deck has against other decks. Let us consider this example card to demonstrate the problem:
This card has a 50% chance to win and a 50% chance to lose the game against any given opponent. It is therefore a balanced card; it can never dominate the meta, it can never be oppressively broken, but it will always, no matter the situation, be boring to play and play against. The reasons why should be obvious: no matter whether you win or lose, neither you nor the other player did anything to influence the outcome of the game; you could have put your pet goldfish in charge of the match and the outcome would be no different.
This problem can be measured by deck polarity. In this case: against half of all decks, this card has a 100% win rate, and against the other half, it has a 0% win rate. This is a very unhealthy place for a card, or deck, to be.
Deck polarity is an objective, measurable way to determine the fun of a deck. A deck that has, for example — 70% win rate vs half of decks, and 30% win rate vs the other half, has less player interaction than one that has 60% and 40%.
A deck which has 90% win rate vs one deck, and 10% win rate versus five equally common decks, is another example. It is balanced; the deck will soon fall out of favour as players begin to play the other decks to counter it, but it will never fun to play or play against.
OTK decks — by their non-interactive nature — always have worse deck polarity than other decks. This can be seen in data such as Vicious Syndicate’s Data Reaper and others. They always perform excellently against slow decks, which cannot defeat them before their combo is assembled, and perform terribly against aggro decks, who can pile on damage from instant effects before the OTK deck has a chance to stall them. It is for this reason that in Hearthstone, control decks that do not utilise OTK combos have not been part of the meta for quite a long time; only sufficiently aggressive decks are capable to stand up to them.
It is also worth noting that as many OTK decks, embracing the one-player strategy, use cards such as Research Project and Coldlight Oracle to draw cards for both players, this amplifies the effect: control decks will overdraw, while aggro decks that would otherwise begin to lose steam will gain extra power.
Here I will demonstrate the above point that only aggro decks and combo decks remain in the meta, using the two most recent Vicious Syndicate Data Reaper reports. I have used the All Ranks reports here, but the Legend rank reports are largely the same in terms of Aggro and OTK deck content.
There are zero control decks listed in either report above. Every single one of these decks is intended to either counter OTK decks (aggro) or to win the game in a one player fashion (OTK). The inability to face OTK decks without fast damage has effectively barred control decks from the game.