Flexibility in Hearthstone’s Arena

This guide covers drafting and playing flexible and situational cards in Arena. Learn how to value flexibility in your draft, decide how many situational cards is too many, and choose when it's best to play both flexible and situational cards.


Gather round the hearth, friends, to hear a tale of an impeccable run at glory in the Arena. The architect of this run was none other than Uther Lightbringer. Contained within the mighty Paladin’s deck were mainstays of the Arena: shielded-minibot, aldor-peacekeeper, piloted-shredder, and dr-boom. But of all the lights in this deck, which one shined the brightest? That, my friends, would be keeper-of-uldaman. Why? Well, that’s the topic of our conversation today: flexibility.

During this run, Boom won me two games and sealed out another. Keeper, though, won me three games and sealed out two. Of the three games Keeper won me, two of them came from upgrading a stealthed worgen-infiltrator and bopping it into an enemy 2/3 minion, basically spending 4 mana on a 3/4 and a 3/1 and a 1-attack boost. The third came when my opponent spent an entire turn playing a big minion that was easily neutralized with Keeper, creating a tempo swing that won the game.

That’s what makes cards like Keeper so good in Arena: they can be used in lots of different situations for different kinds of effects. In other words, they’re flexible. But playing cards like Keeper can be difficult, since they require a keen eye for the current game state and their potential effects.

This guide covers both flexible and situational cards, including what to think about when you’re drafting them and how to decide when to play them. It begins by covering terminology, then covers how to draft flexible cards, play flexible cards, draft situational cards, and play situational cards.


Keeper of Uldaman is one of the classic examples of a flexible card in Hearthstone: it can be used in almost any game situation, and it can be used in many different ways: as single-target removal, as a silver-hand-recruit buff, or even as a body on an empty board.

Another Keeper, keeper-of-the-grove, is similarly flexible. Both 2 damage and Silence are relevant effects, and they are relevant in different game situations. In fact, the Choose One mechanic for Druids means that Druid is perhaps the most flexible class, at least in terms of minions.

Another example: seal-of-champions is often better than blessing-of-kings because of flexibility: the times when Kings excels can be similar to Champions (mid-game minion-on-minion combat), but Champions works better sending small minions into large minions and when sending most minions into the opponent’s face.

Other flexible cards that fit this definition include polymorph-boar, which can serve as either a single-target removal or 4-damage burst, and madder-bomber, which can be played for a last point or two of face damage, as an arcane-explosion-style weak board clear, or to finish off a single enemy minion at lower health. In other words, flexible cards are ones that can be used in several different situations, especially common situations.

Another way to think about flexibility in a card is to consider when in the game the card is good. For instance, consider bloodfen-raptor. It’s a solid card when played on Turn 2 or Turn 3. It can be functional in the mid-game, played alongside another small card or something like a chillwind-yeti. However, late in the game, it is generally bad.

Compare Bloodfen Raptor, though, to fallen-hero. Like Raptor, Fallen Hero is at least solid on Turns 2 and 3, but it is also very often the best Turn 4 play. It also holds relevance in the late game. It is unlikely to win or even swing the game in later turns, but it can have a meaningful effect in a tight late game.

It’s this kind of curve-related flexibility that makes many Inspire cards great in Arena. boneguard-lieutenant is often just a Bloodfen Raptor, but he can also be buffed in the early game to eat up a 2/3 minion or played in the late game as something more akin to a spider-tank. Similarly, murloc-knight is playable on Turn 4, but it is even better played on Turn 6 or later.

This kind of flexibility is also what makes Discover both a fun and a strong mechanic: it allows you to select a card that works for your current game situation or an anticipated situation, but also allows you to pick cards that can fill holes in your deck or smooth out your mana curve based on what’s currently in your hand.

Situational cards, on the other hand, are cards that are generally only good in either very few or very specific (i.e., not common) situations. Consider a card like molten-giant: the time it shines is when you are either low enough on health for it to be playable but not low enough for it to make no impact on the game, or when you have a ton of healing in your deck. Unfortunately, these situations don’t arise particularly often, making Molten Giant a mediocre Arena card. In other words, much like doomsayer, its window of playability is narrow.

Other situational cards include board buffs like bloodlust/savage-roar and cards with specific triggers, such as grim-patron, kezan-mystic, or gadgetzan-auctioneer. Cards that can significantly harm your own board development, such as abomination are also situational, as they’re generally only usable when you’re behind on the board.

Flexible and situational cards are different from cards that are always good or always bad. Of course, there are no true always in Hearthstone, but the other cards I mentioned in that 12-0 Paladin deck—Shielded Minibot, Aldor Peacekeeper, Piloted Shredder, and Dr. Boom—are pretty much always relevant cards. What differentiates these always good cards from flexible cards is that they do one thing and one thing only, even if they do that thing very well. On the flip side, a card like silverback-patriarch will almost always be a bad card, even if it does save you from certain death on rare occasions.

Drafting Flexible Cards

I’ll begin the strategy portion of this guide by talking about flexible cards, and then I’ll address situational cards. In both cases, I’ll start with drafting, then cover playing.

When you’re choosing between three cards in Arena, the first thing you should do is evaluate a card’s overall value. There are other guides on this site that can help you do that, though you’ll notice that the flexible cards I mentioned above are almost universally strong cards. While there is not a direct relationship between flexibility and card strength, you can generally assume that more flexible cards will be good choices.

Strong Arena players, though, will often make very fast judgments about overall deck quality, the deck’s win condition(s), the gaps in the deck, and mana curve. These judgments should meaningfully change your drafting choices, and one way beginning and intermediate Arena players can get better is by slowing down that judgment process and making it conscious.

So, what should you consider when drafting flexible cards? First, you should value flexible cards more when your curve is suboptimal. As I mentioned in the previous section, flexible cards allow you to change when they are played based on what works for your current mana curve. The more suboptimal your mana curve is, the more times you will find yourself needing to fill holes in that curve, which is something flexible cards often excel at. For instance, a Murloc Knight can both shore up your 4-drop slot and make up for a lack of late game cards simultaneously. If there are weaknesses in both your mid-game and late-game curve, a Murloc Knight is generally a stronger choice than a card that solely dominates the late game, such as north-sea-kraken.

Another example: dark-peddler can fill quite a few gaps in your curve. If you have fewer 2-drops than you would like, it can help out there. The same goes with 3-drops. It can also serve as a mid-game card, where a 1-mana spell might be very relevant.

Dark Peddler also demonstrates another way to think about drafting flexible cards. You should always consider the roles different cards fill in your Arena deck. In general, the roles cards can fill are as follows:

  • Solid minions (i.e., simple “drops” at each mana slot up to at least 4), such as shielded-minibot
  • Board clears, such as consecration
  • Small removals (single-target), such as frostbolt
  • Large removals (single-target), such as flame-lance
  • Pings, such as mortal-coil
  • Reach, such as crackle
  • Healing/Taunts, such as tuskarr-jouster
  • Card draw, such as battle-rage
  • Late-game finishers/engines (i.e., cards that can help you seal out the game), such as kodorider

Not all those roles are absolutely necessary for every Arena deck, and good Arena players know that different win conditions mean increased or decreased importance on different roles (an obvious example is that healing and Taunt cards are less important in tempo/aggressive decks than in value/control decks).

However, what flexible cards allow you to do is shore up weaknesses in a deck without having to dedicate an entire slot to them. Take Keeper of Uldaman: there are plenty of games where a large removal is not particularly necessary or relevant, and in those games, you can use Keeper to buff your own minion or on your opponent’s mid-range minion, where an equality would just sit dead in your hand.

Finally, after you have chosen a flexible card, you should adjust your drafting strategies accordingly. This means not only prioritizing cards that fill other roles, but it also means identifying the best case scenario for your flexible card(s) and drafting in a way that makes those scenarios come up more often. The Worgen Infiltrator into Keeper of Uldaman play I mentioned earlier in this article was actually one that I saw potential for when I draft Worgen Infiltrator. Similarly, if you draft a Madder Bomber, you might steer away from low-health minions. These kinds of things shouldn’t cause you to dramatically shift your draft, but they should at least be considerations. Even if a particular combo or situation only appears a few times in a run, that’s a meaningful thing that can change your win percentage by a few points.

Playing Flexible Cards

The previous section is only something you should be concerned about when you’re faced with a difficult decision between a strong standlone card and a more flexible card. However, this section deals with decisions you will encounter in almost every Arena game: when should you play the flexible cards in your hand and when should you keep them for later? By their very definition, flexible cards can be played at a lot of different points, and it can be tricky to identify the right time to play a flexible card.

There’s a concept in sabermetrics—the statistical study of baseball—that sheds light on this issue: Win Probability Added (WPA). The statistic measures how much any action a batter or pitcher takes adds or subtracts from his team’s likelihood to win the game. A grand slam in a tight game will make the batter’s WPA go up quite a bit. A double play in that same situation will make his WPA go down, as his team is less likely to win afterward than they were before the double play.

When you play Hearthstone, basically every turn you should be making quick WPA calculations in your head, asking yourself: what plays are most likely to increase my chances to win this game? This means thinking ahead to how your opponent is likely to play their next few turns, as well as what game situations are likely to arise later in the game and how you might deal with those situations. This also requires an understanding of how your deck is likely to win games, since those calculations should be skewed heavily toward your win condition(s).

A good rule of thumb when it comes to WPA calculations is: if you can appreciably increase your likelihood of winning the game with a single play or series of plays, do it. Arena games, like baseball, are often games of inches, and if you find yourself waiting for the massive swing play to come along, then you may well wait your way into a loss.

Going back to Keeper of Uldaman as an example, there are 5 general situations when you can play Keeper:

  • On an empty board or on a board where the buff/debuff doesn’t matter, such as facing an enemy earthen-ring-farseer
  • On a big enemy minion, such as boulderfist-ogre
  • On a midrange enemy minion, such as chillwind-yeti
  • On your own small minion, such as silver-hand-recruit, played that turn
  • On your own small minion played previously and able to attack

Playing Keeper on a big enemy minion or your own small minion played previously and ready to attack represents a massive WPA swing (see the Worgen Infiltrator example above). However, playing Keeper on a midrange enemy minion or on Turn 6 with a newly-summoned Silver Hand Recruit represents an appreciable bump in WPA. Spending 4 mana on a 3/4 minion and dealing 2 damage to the Yeti or spending 6 mana on a 3/3 minion and a 3/4 minion are both reasonable investments of mana that will make the game state swing in your favor. You shouldn’t ignore these options just because they don’t represent the kinds of massive swings that other options do.

That said, you also shouldn’t necessarily take these options just because they’re available to you. Let’s say you’re considering whether or not to hero power + Keeper on Turn 6. If your other options are very mana inefficient, then this is likely the best play. However, if you have a silver-hand-knight in hand, you can get 1 less stat point on the board but save the flexibility of using Keeper for removal of an enemy minion. By this logic, if you have two cards that can do the same thing, you should save the flexible card, since it can do more than that thing.

This leads me to a concept I like to refer to as “card equivalencies.” There are a number of situations where two cards are likely to achieve roughly or exactly the same thing. A few examples:

  • Your opponent plays a Turn 2 bloodfen-raptor. You have your own Raptor and a frostbolt in hand. Both cards are likely to end in an empty board for your opponent’s Turn 3 play.
  • Your opponent plays a Turn 2 Raptor. You went first, and you can play either harvest-golem or argent-horserider Turn 3. Both cards are likely to end with you having a 2/1 minion for your opponent’s Turn 3 play.
  • Your opponent has a 2-health minion on Turn 6. You have argent-commander and stormpike-commando in hand. Both plays are basically identical.

At times like these, the general rule is that you should play the less good card (e.g., Bloodfen Raptor over Frostbolt), since the better card is likely going to be more useful later in the game in another situation. This applies even more to flexible cards. When another card is likely to achieve the same goal or increase your WPA the same amount, you should use that card before the flexible card. For example, you should almost always fireball + ping a boulderfist-ogre rather than polymorph + ping, since the Polymorph might be useful later to silence something or deal with something with more than 7 health (unless you need the Fireball for reach).

In many ways, this advice mirrors my suggestions in my article about single-target removal in Arena. Don’t be greedy and be willing to play a flexible card at a time that is not the absolute best that card can do, since that perfect time may never come. At the same time, recognize when another card will accomplish roughly the same thing and be willing to hold onto the flexible card at those times.

Drafting Situational Cards

Flexible cards are ultimately pretty easy to draft (though a little less easy to play). Situational cards, on the other hand, are much trickier to draft. Part of this is because situational cards are so variable in value. Flexible cards and always good/always bad cards can be valued pretty consistently, though your current deck composition should affect those values slightly. However, the value of any given situational card is heavily dependent on the deck surrounding it. A summoning-stone or gadgetzan-auctioneer on the 20th pick with 2 spells so far is pretty bad. Those same cards on the 20th pick with 8 spells are good, even great cards. I recently drafted a Rogue deck with an assassins-blade, a deadly-poison, 2 tinkers-sharpsword-oil, and 2 blade-flurry, and those Blade Flurry performed better than I have ever seen Blade Flurry perform in Arena.

The best thing to do regarding situational cards, then, is to identify the situations where the card is good, then weight your deck to make those situations arise more often. If you pick abomination early, then weight your deck to be a little slower than usual, so Abomination will be good more often than he would otherwise. If you pick divine-favor early, then weight your deck to be a little faster than usual so you can trigger Divine Favor more often.

Sometimes these deck tweaks aren’t as obvious as “pick up a bunch of spells to make Gadgetzan Auctioneer work better.” Auctioneer is at its best with a few Taunts to hide behind and smaller spells, especially removal spells that can help clear the board for multiple turns of Auctioneer activations.

The real trick here, though, is to not go overboard with making your deck fit your situational cards. Your deck still needs to be balanced and successful regardless of the situational card(s). Remember that if you have only one copy of a card in your deck, you are only going to draw it in half or less than half of your Arena games, and you are only going to have it in your opening hand about a quarter of the time (even with an aggressive mulligan). In other words, try to think about how the game is likely to play out if you don’t draw that situational card you weighted your deck around. If you go too crazy with a low curve to make Divine Favor better, you may simply run out of steam and lose before you draw Divine Favor.

An exception can be made here when weighting your deck to make your situational cards better doesn’t really sacrifice anything. For example, like Divine Favor, battle-rage is a very situational card draw, but unlike Divine Favor, the things that help trigger Battle Rage—deaths-bite, cruel-taskmaster, etc.—are independently strong. Similarly, Savage Roar and Bloodlust might suggest a heavily aggressive deck, but those decks can often win without those cards.

Let’s take a closer look at this idea with a specific card: perhaps the master of all situational cards, and my own personal favorite card in Hearthstone, hobgoblin. Hobgoblin’s value as a card is heavily tied to the number of 1-attack minions in your deck. Have enough 1-attack minions to make him more than just a 3-mana 2/3 and you risk having your opponent just steamroller your weak minions. That’s why I only ever really pick Hobgoblin in Warlock, because the hero power compensates for having weak minions.

In fact, I recently drafted two different Hobgoblin Warlock decks in a two-week span, and I think they illustrate this balance of adjusting your deck to accommodate a situational card without overcommitting to that situational card. Let’s take a look at those decks:

Deck 1 ended up going 7 wins and Deck 2 ended up going 11 wins. Some of that may be variance, but there’s also something about the deck composition that made the difference here. Amazingly, both decks had the same wombo combo of murloc-tinyfin, wisp, and Hobgoblin. When I drew Hobgoblin, I was able to play them together for a massive tempo gain. When I didn’t, I usually saved Wisp and Murloc Wisp for the times when they were strategically valuable.

However, looking at the first deck, there are a number of 1-attack minions that aren’t particularly good: cogmaster, undertaker, lance-carrier, micro-machine, mogushan-warden. None of these cards are completely terrible (well, okay, Mogu’shan is pretty bad), but they aren’t good enough to sustain the deck on their own, either. In essence, I tried to surround my situational card with other situational cards, a strategy that doesn’t usually pay off. Despite the 2 imp-gang-boss, Deck 1 just pretty much lost when I didn’t draw Hobgoblin and won when I did.

Compare that to Deck 2. The 1-attack minions here are much better on their own: voidwalker and annoy-o-tron are really solid small Taunts, haunted-creeper is perhaps the best neutral 2-drop in the game, and jeeves, while situational, is perfectly suited for the same deck as Hobgoblin (lots of small cards). Only master-swordsmith stands out as a bad card without Hobgoblin. Granted, this deck also filled more of the roles listed in the Drafting Flexible Cards section (siphon-soul for large removal, antique-healbot for healing, etc.), but the inclusion of solid 1-attack minions here meant that the deck won plenty of games without Hobgoblin.

One final note: some situational cards are much less situational in certain classes. molten-giant is much more likely to be playable in Rogue and Warlock than in other classes. jeeves is great in Hunter and Warlock, where your curve is generally lower by default. When drafting situational cards, especially early in your draft, consider how that class’s typical play style works with the conditions of the card.

Playing Situational Cards

When it comes to playing situational cards, my advice is basically just a more drastic version of my advice in the Playing Flexible Cards section: be willing to play these cards at moments that aren’t perfect. In fact, acceptable is not a bad bar to hit when playing situational cards. If you find yourself looking at one of your situational cards in your hand and saying, “meh, that’s not half bad here,” you should probably play it, because waiting until it’s good is often just a pipe dream.

The general idea here is that when you play a bad card, you improve the overall quality of your hand. As I said in my article about drafting and playing terrible Arena cards, if you can identify a turn when a bad/situational card won’t hurt your chances too much (i.e., minimize the WPA loss), then you can dump that card and make your future turns stronger as a result.

Another technique for playing situational cards is to identify when the card is solid enough without its effect. Take Gadgetzan Auctioneer again. Its effect is big enough that you should generally not be playing the card without being able to trigger its effect. However, if your opponent’s board consists only of an ancient-shade, an empty Auctioneer is actually a reasonable play. Similarly, Jeeves with a large hand and into an opponent’s Yeti is a complete disaster. Jeeves with a large hand and into an opponent’s 4 imp-losion imps is mostly fine.


I hope this article has given you insights into how to both draft and play situational and flexible cards in Arena. My next Arena strategy guide is going to be about making the late game push as an aggressive or tempo deck, including how many big cards you should draft, how to value card draw, and how to ensure you kill your opponent before your deck runs out of steam. I’m also really open to writing about any other Arena-related topics you all have an interest in, so if there’s anything you’d like to see me cover, drop me a line in the comments section below, and I’ll try to take a stab at it. Thanks for reading, and happy Hearthstoning!