Preparation, Strategy, Tactics, and Action: Phases of Challenge

Many games are tests of skill. Players succeed or fail at the game’s goals based on their physical dexterity and reaction time, general knowledge and reasoning ability, understanding and internalization of the game’s own mechanics - anything a game can test. But much of that skill is applied before the moment of success or failure.

Victory in a chess match may come from physically moving your piece into a position that checkmates your opponent, but that isn’t the hard part. And the hard part of beating Doom isn’t the button press that fires the last shot on the final boss - it’s everything you did to enable that shot. These goals, and indeed most interesting goals in games, actually have multiple stages of challenge that funnel into each other.


Here’s my conception of the phases of challenge. This is a fairly abstract framework, since it’s intended to be generalizable to every skill-based game. To help pin it down a bit, let’s take a closer look at each phase and then discuss how they interrelate. Once that’s done, I’ll go into some implications these ideas have for game design.

Defining the Phases


Preparation is about getting ready to deal with the game’s challenges.

There are two types of preparation. Player preparation is about the player themselves improving their own skills. This may be through research to learn, understand, and internalize the game’s systems and iconography or through practice of the abilities it tests.

This kind of preparation can be done separately from any engagement with the game’s defined goals, and sometimes without directly playing the game at all. Because it affects only the player, the improvements here will follow that player between playthroughs and to other similar games, but will not benefit any other players using a shared account, save file, or equivalent.

By contrast, in-game preparation is about improving the ability of the player character (or equivalent) to deal with the game’s challenges. This is done through acquiring new options (getting new abilities, unlocking doors, etc.) or improving existing options (upgrading a weapon, boosting character stats, etc.). Different kinds of games have different approaches to this, but some are standard for certain genres. Common ones include progression through certain points in a game’s story (for example, rescue engineer to receive airship), accomplishing specific optional objectives (complete courier side missions to improve running speed), grinding or farming (gain experience levels to learn new spells), defeating other players (win PVP battles to earn armor), spending an in-game resource (buy potions with in-game gold), or spending real money (buy card packs with real-world money).

As this preparation occurs within the game’s universe, it requires interacting with the game in some manner. Additionally, the effect is internal to the game, so it won’t follow the player between playthroughs or to similar games unless specifically accounted for (see New Game Plus, Old Save Bonus). However, it does transfer between players using a shared account, save file, or equivalent.


The next phase is strategy. This phase is about defining a framework for handling a game’s challenges.

There are two distinct but interrelated aspects of strategy. A plan is a multi-step process for overcoming a challenge, defined in advance. Depending on how predictable the challenge is, the plan can be rigid and deterministic (jump across these platforms, flip the switch, shoot the security drone, then walk through the exit) or can include reactions to hypothetical events (defend this room, turning to whichever door an enemy comes through and using an attack of the particular enemy’s elemental weakness - unless several enemies enter at once, in which case retreat to the safe room and heal up). Because a plan exists in the mind of the player, it can be changed on the fly in response to evolving circumstances.

Meanwhile, a loadout is the specific set of options that will be usable when dealing with a challenge, chosen as a subset of all available options. This can take many forms such as a character build, an equipped set of gear and abilities, a battle party, or a card deck. Loadouts can be multi-leveled - for example, you may put together a battle party by selecting from many characters and then determine each character’s equipment and abilities. Loadouts exist within the game world, so opportunities to change them vary from game to game - but it’s often the case that changing loadouts is slow, requires spending in-game resources, or is only possible during downtime. In many cases, you are locked in to a particular loadout for the duration of whatever constitutes an individual test of its efficacy - for example, if it’s a combat-related loadout, it may be difficult or impossible to change the loadout during battle.

A plan and a loadout influence each other. If particularly useful options have become available, the player may define a loadout around them and then plan based on this loadout. Alternately, a player may first define a plan and then choose a loadout that enables this plan.


The third phase is tactics. This phase is about making choices in response to specific situations.

There are two sides to this. Awareness is the understanding of the game’s current situation and all factors in play - the “game state.” In order to choose a response to a situation, you have to know what the situation actually is. Games can vary considerably in how much of the game state is clear to the player. Information may be withheld completely (it’s unknown what the enemy units are doing behind the fog of war), provided vaguely (the edges of the screen may become bloody to indicate you are wounded, but not indicate precisely how much health you have left), or given as a probability rather than a certainty (the shot you are considering taking has a 67% chance to land successfully). Whatever information is provided will be filtered through the game’s iconography (a boss might switch from a walking animation to a limping one to show that its health is somewhere below 25%). In some cases, it may require memory or reasoning on the part of the player (a countdown was started but is no longer visible; an ability was used that will have an effect in four turns).

Decision is choosing which of the options in the current loadout to use in response to the current game state. (Use a stealth takedown on the unaware guard, have Pikachu use Thunderbolt, select the “sarcastic” dialog option, etc.) Naturally, this depends a great deal on awareness so that the correct game state is taken into account. Otherwise, the decision can backfire. (The guard actually is aware, Pikachu’s fighting a Ground-type, the person you’re talking to is hostile, etc.)


The fourth and final phase is action. This is the process of transmitting a decision back into the game to update the game state. This is often the moment of truth - whatever skill is being tested by a particular challenge, this when the player either succeeds or fails. This is true even if most of the work was done in prior phases.

In some games, which I will broadly refer to as “action games”, this is expected to be part of the challenge and as such requires nontrivial amounts of physical capability (aiming a headshot at a moving target, hitting the right button sequence to execute a combo move). In non-action games, this is not expected to be part of the challenge and is merely its interface. As such, action in these games is designed to be trivial for able-bodied people (selecting an option from a menu, moving a chess piece, revealing a card).

While both kinds of games can benefit from accessibility accommodations that make action easier for people with disabilities, the situation is more complicated if the game is competitive. In action games, accessibility concessions are often seen as disrupting competitive balance while in other games they are seen as restoring it.

How the Phases Interact

I’ve presented the phases linearly. This is because they do tend to come in this order for a given challenge, with each phase constraining and flowing into the later ones.

Preparation constrains strategy. Strategy is about defining a framework to make use of the advantages gained in preparation. In particular, player preparation constrains planning, as a plan can only make effective use of a game’s systems to the extent that the player understands those systems and possesses the required skills. (You can only plan to make use of enemy weaknesses if you know what they are and can aim for them.) Similarly, in-game preparation constrains loadouts, as a loadout can only include options that have become available within the game. (You can’t include an unrecruited character in your party, equip a weapon you haven’t yet bought, or include a card in your deck that you haven’t acquired.)

Strategy constrains tactics. Tactics are about choosing how to respond to individual situations in ways consistent with the defined strategy. Loadout constrains what decisions are possible, as you can only choose from your usable options. (You can’t use your ice sword against the fire demon if you didn’t bring it with you.) Additionally, the quality of a decision is constrained by the plan, as a plan will not succeed if individual decisions do not advance it. (You could choose to shoot your ally instead of your enemy, but this won’t be helpful if your plan is to defeat your enemy and protect your ally.) The plan also defines what specific elements are essential for awareness - if the plan requires responding to certain events, you must notice when they happen or the plan will fail.

Tactics constrains action. Action is about applying a tactical choice back into the game. The chosen option fully defines what action must be taken within the game’s systems. (Deciding to climb to a higher vantage point means you must apply the right inputs to accomplish that climb.)

But in practice, causality can flow in both directions. It’s possible to start with preparation, devise a strategy based on the gained advantages, make appropriate tactical choices and take the corresponding actions, but it can go the other way too. Perhaps you enjoy a particular action, such as firing a shotgun that shoots lightning. You want to be able to frequently use it as an effective tactic, so you design a strategy around it and then do the required preparation to enable that strategy.

Additionally, in most games individual challenges feed into each other in one or more gameplay loops. The results and feedback from an action often go right back into tactical awareness. (If you killed the enemy you’re attacking, move on to the next; otherwise, stay on this one.) Preparation is also often essentially just repeated action. Each attempt constitutes practice, and the outcomes may constitute research. (You may get increasingly good at headshots, but then observe that a particular enemy type is actually better handled by attacking its limbs first.) In-game preparation requires action as well - sometimes one-shot (you picked up the red keycard and can now open red doors) and sometimes repeated (each battle you win gains you experience which raises your level and improves your attack and defense stats).

However, the time frames generally shrink as you move forward through the phases - both in terms of the time available to work within a phase and the duration of the effect of doing so. Preparation is often unconstrained and permanent. Strategy may take a few minutes and have effects lasting several minutes to multiple hours. Tactics may be on the scale of a minute or two, while actions might be immediate and have direct effects that only last a few seconds. The actual scales vary between challenges and between games.

Finally, it’s worth noting that different games (and different challenges within a game) distribute challenge differently between phases. For example, chess has no real concept of in-game preparation due to every match starting with the same state and the same pieces, while poker does have in-game preparation as the size of your stack influences what options are available - you can only bet or call with chips you actually have. We can refer to the distribution of difficulty across phases for a particular game or goal as its challenge profile.

Implications for Design

I see several game design takeaways from this framework. Here they are.

1. Challenge profiles should be clear.

Different people enjoy different challenge profiles to varying degrees. One person may enjoy preparation due to the feeling of continual improvement, another might consider it an obstacle that just delays wide-ranging strategic experimentation, a third might find that too slow and prefer the quick feedback of tactical decision-making, while a fourth might focus on the flow-inducing immediacy of action. And of course an individual’s tastes can change with their mood - a player might enjoy action games earlier in the day but then want to relax in the evening with some grind-heavy preparation.

As such, it’s useful for players to have a good idea in advance what phases of challenge will be important in a given game. This is one of the major benefits to genre labels, as many genres imply a particular range of challenge profiles. A player who dislikes action-based challenge would do well to avoid first-person shooters, while one who dislikes in-game preparation should avoid idle games. Some genres even have a dominant phase right in the name, such as “real-time strategy”, “tactical RPG”, or “action adventure.”

Until Dawn cover art
Games that don’t fit in well to established genres are thus at greater risk of having challenge profiles that surprise the player. For example, the difficult-to-categorize Until Dawn bills itself as a game of decisions. Its official website mentions five times that your choices will determine the story, and three of those mentions specifically say that your choices determine who lives and who dies. By contrast, the page mentions quick-time events zero times. But in practice, it turns out that your decisions don’t really matter if you can’t pass the QTEs. This can make for a very frustrating experience for players with slower reaction times who came to play the promised game of decisions and weren’t looking for action-based challenge. Deciding what characters should do in the moment (an interesting tactical challenge) is superseded by QTEs (an uninteresting action challenge).

It’s also important for an individual challenge within a game to make its profile obvious. It’s commonly understood that games must provide clear feedback to the player on the results of their actions in order for the player to effectively learn and improve. But it’s very easy to make the mistake of providing action-level feedback to challenges that aren’t really about action. This is especially true because different challenge phases tend to operate on different time scales - it’s much easier to get immediate feedback for actions than for strategies, for example. But it’s absolutely worth mitigating this issue when possible. A player who doesn’t know what they’re doing wrong is not in a good position to learn to do better.

Suppose, for example, you are playing a Metroidvania and there is a ledge that you can almost jump to. You try the jump a couple of times and just barely miss each time. Does this mean you are timing your jump poorly and you need to improve your action? Does it mean jumping is not the way to get to this ledge and you need to choose a different tactic? Does it mean you aren’t supposed to go to the ledge at all and this is the incorrect strategy? Or does it mean you need to prepare first by finding a jump upgrade power? If the game provides no other cues, it can be quite hard to tell. You might waste time and frustrate yourself by retrying the jump several times to get the timing just right when that’s not the problem at all. If instead the platform were further out of reach and it was clear your jump was nowhere near enough to get to it, you’d be much more likely to quickly realize that’s not the right approach.

2. Optional trade-offs between phases allow for more playstyles.

If different people enjoy different challenge profiles, then having more high-difficulty phases of challenge in a given game will shrink a game’s potential audience. Players who don’t enjoy preparation, strategy, and tactics are unlikely to persist through a game that requires high amounts of all three.

But that’s only if the profile is inflexible and you genuinely need to work hard in each of the phases. If instead it’s possible to invest heavily in one phase to offset the requirement of another phase, that means the game supports multiple playstyles and the potential audience becomes larger - more players of different tastes will be able to play the game in a way that appeals to them.

I once wrote about the fact that many RPGs present battles as challenges that can be overcome either through strategy and tactics or through preparation - you can use your resources optimally, or you can grind levels and get enough resources that you can blatantly misuse them and still win. This means that those who enjoy the strategy and tactics challenge can experience it, while others can get around it.

Taking it further, this kind of flexibility enables people to concentrate difficulty into particular phases via self-imposed challenges. Players can eschew level-grinding preparation in RPGs and go hardcore on the strategy and tactics challenge via a low-level run. Similarly, Metroidvania games often provide a lot of preparation in the form of upgrade pickups, but a player can instead concentrate the challenge into the action phase via a minimalist run.

3. Punishment should be scoped to the phase where the player needs to improve.

Punishment is what happens when the player fails. For it to feel fair to the player and be useful for learning and improving skills, it should be scoped to the particular challenge the player has failed. (Defending this claim is beyond the scope of this article, but see Test Skills, Not Patience: Challenge, Punishment, and Learning for related thoughts.) If that’s the case, then that suggests that punishment should also be scoped to the specific phase or phases where the player failed.

Suppose you are playing an action RPG. You attempt to attack an enemy, but your sword swing misses due to bad timing. This is an action-level failure, and the punishment is kept at the action level - you deal no damage to the enemy but you can you freely try again. You attack again, and land the hit - but realize the enemy is actually blocking physical attacks and you need to use magic instead. This is a tactical-level failure, and the punishment is kept at the tactical level - you deal no damage to the enemy, but you can pay attention to the enemy’s stance and choose a different approach.

So you lob a fireball, but then realize that this enemy is resistant to fire damage. You poured all your skill points into fire magic, but enough enemies resist fire that you probably should have diversified into multiple elements. This is a strategy-level failure - is the punishment kept at the strategy level?

Torchlight 2 respec potion
Only if the game allows you to “respec” - in other words, only if you can refund and redistribute your skill points. If you can’t, then your punishment is actually on the preparation level - to fix the situation, you need to earn more skill points. Depending on the situation (How frequently are skill points awarded? Is it prohibitively difficult to deal with the current enemies with only fire magic? Can you get useful amounts of progress toward skill points by fighting lower-level enemies or enemies in a different location?) you may even have to start over with a brand new character to do this effectively, rendering moot all the in-game preparation you’ve done so far.

But if you can respec - that is, if you can refund and redistribute your earned skill points - then your punishment is indeed on the strategy level. You can redo your loadout to include some ice magic and be on your way. A respec mechanic makes in-game preparation exchangeable - you can change your investment, but you don’t get more to invest than you’ve already earned. As such, it doesn’t undermine preparation, but it does allow the player to experiment and learn on the strategy level without forcing them to also repeat large amounts of preparation.

If you want players to improve their skills in a certain phase, scope the punishment to that phase. It’s unfair and unproductive to punish them in phases where they have committed no errors.

4. Unpredictability pushes challenge back a phase.

Suppose you are playing a first-person shooter in which your character has a highly-accurate pistol. In such a game, your ability to successfully shoot enemies - an action challenge - is down to your own skill at that very action. But suppose your character’s gun were less accurate, with some amount of random scattering of its shots away from where it’s aimed. There’s still action in the challenge, but your success is determined less by your skill at aiming and firing. Now tactics enter into the challenge more - you have to know when to fire, bringing the enemy close enough for the randomness not to matter much.

Alternately, consider Hearthstone, a collectible card game in which you build a deck from your owned cards and then use that deck in a match. Choosing which card to play on a given turn is a tactical challenge; however, you don’t necessarily know which cards you’ll have in hand at any given time. Your deck is shuffled before a match and in most cases you only draw one card per turn. This loads more of the challenge into the strategic level of building the deck - you have to account for the unpredictability and try to make a deck that can function regardless of the order in which you draw the cards.

Or consider Pokémon. Putting together a competitively-viable team is a strategic challenge. But once you’ve decided which ‘mons to include and caught one of each, you’re not done - Pokémon have some random stats that affect their capabilities. As a result, you have to do a lot more preparation to actually build your team (probably involving breeding).

If a player can’t predict or control the outcome of a particular phase of challenge, this pushes more of the challenge into the previous phase to compensate. This isn’t automatically a problem, but it’s definitely something to be aware of when designing a game and aiming for a particular challenge profile.

5. Reducing one challenge phase can emphasize another challenge phase.

I mentioned above that reducing action challenge via accessibility accommodations can be seen as restoring competitive balance in non-action games. This is because - by definition - the action phase is not important in non-action games. If you were playing chess with someone who was unable to move their own pieces and used voice recognition software to make moves by saying things like “Knight to Rook Five,” you would likely still consider this a fair chess match. If instead you were playing Team Fortress 2 against this person and they could take action by saying things like “Run to the capture point” and “Headshot the enemy scout,” this would feel less fair - the two of you aren’t really playing the same game because your opponent has a very different experience of the action phase.

If you were both using the voice-activated version, things would be fair again, but you still wouldn’t exactly be playing Team Fortress 2. You’d be playing a variant in which action is de-emphasized and strategy and tactics are emphasized - your strategic and tactical skills have an increased effect on your odds of victory. Thus, this version of Team Fortress 2 will appeal more to players who enjoy strategy and tactics over action challenges.

Stephen's Sausage Roll screenshot
As such, game designers may deliberately reduce challenge in one phase to concentrate it in other phases. This is often a good way to prevent things from becoming tedious by ensuring that what’s difficult about a game is also what’s interesting about it. For example, in his review of Stephen’s Sausage Roll, Joseph Anderson notes that the game’s unlimited undo feature allows you to solve puzzles while figuring them out through active experimentation, rather than thinking them through without acting and then being in the position of knowing what to do but still having to do it. In other words, uninteresting action challenge is drastically reduced in order to emphasize the far more interesting planning and tactics challenges.

But if this sort of refocusing comes as a change to an existing formula, such as in a patch or a sequel, it’s often controversial. Some number of players are likely to be invested in the prior challenge profile. Such a player is likely to see this change not as streamlining or fat-trimming, but dumbing down - especially if they are proud of their achievements in the now-diminished challenge phase. See for example the mixed response to Blizzard mostly removing instance attunements from World of Warcraft, allowing players to participate in raids (strategy, tactics, and action challenges) without having to do specific lengthy quest chains first (preparation challenge). Some players celebrated the game removing a feature that “existed to keep you from doing what you wanted to do until you’d done what you didn’t” while others lamented that this “diminished the value of raiding, and diminished the value of guilds.

Wrapping Up


The challenge phases as I’ve defined them are just one way to categorize and think about challenge in games, but I think it’s a useful one and it’s already helped me have more productive conversations about game difficulty. The design implications listed above are mostly based on ideas that I’ve had half-formed for some time. This framework has finally allowed me to crystallize and express them.

Challenge is one of the most important topics in game design, yet it’s commonly misunderstood and oversimplified - we say a game is too hard or too easy when what’s really going on is far more complex and interesting. Proper understanding of the elements of challenge is crucial to designing games that use it well.