Meditations on Strategy #2: Feedback

Now that I’ve talked about theory, it’s time to talk about problems and solutions.

After making sure the game has the potential for strategic depth in the first place, the most crucial element in a strategy game (in my opinion) is feedback. This is a concept with several names; Rules of Play refers to it as a type of “external event” (those concerned with representation of choices to players, as contrasted with “internal events” that relate to the choice itself), and Soren Johnson in his Art of Strategy talk calls it “transparency.”

Put simply, feedback refers to the manner in which the game conveys the potential consequences of a choice (and, in some cases, the ability to make the choice in the first place) to the player. Ideally, feedback should be completely intuitive, so that a player fully understands the consequence of every choice before making it. For instance, in the FPS genre, it can be taken for granted that shooting an enemy will injure or kill it, shooting it in the head will injure or kill it far more quickly, and missing will not injure or kill it. This is a kind of feedback that generally requires no special consideration on the part of the game designers (though FPS games in which enemies can take a great deal of hits before dying, such as Borderlands, often result in complaints about ‘bullet sponges’—a breakdown in intuitive feedback when enemies don’t take damage from being shot as quickly as players expect them to.)

As well as being an impossible ideal in games of a more abstract and less realistic sort, perfectly intuitive feedback may at first seem at odds with the definition of a strategy game I put forth in my last blog entry. Didn’t I say that “obvious choices” lack strategic value? Isn’t a choice with perfectly intuitive feedback—such as the decision to try and shoot an enemy in the head—not strategic? The key to understanding why this is not the case is that strategy in a game with effective feedback evolves not due to individual choices themselves, but due to the interactions of them with other choices and gameplay elements.

Let’s look at the new XCOM again. The cover mechanic is a great example of intuitive feedback; full cover is safer than half cover, which is safer than no cover, and the player quickly comes to understand this as a result of the tutorial if they don’t upon sight. (This breaks down somewhat in the game’s highest difficulty, where the only reliable way to avoid getting hit is to distrust all cover and stay out of the enemy line of sight entirely, but performs well as a rule of thumb on normal difficulty.)

A player character may move into full cover and hunker down, seemingly ensuring it’s all but invulnerable to enemy fire—an ‘obvious choice’ if total defense is the player’s objective, such as when controlling a civilian being escorted to safety. But other factors may reduce or eliminate the safety of this ideal position. If the player has allowed the enemy to get too close, an alien may flank the character from a direction it has no cover from, making it far more vulnerable. An alien may use an explosive or other attack that ignores cover. Tactical considerations, such as the position of the enemies and the type of enemy (do they have cover-ignoring attacks?) heavily influence the ‘obviousness’ of the choice.

Alternately, an alien may beat the odds and land an unlikely shot on the well-protected character, or miss but succeed in destroying its cover. Randomness is a special type of factor in strategic choice that can be displayed to the player to achieve theoretically ‘perfect’ feedback, but is not necessarily satisfying for them. I’ll discuss it in more detail in the next blog entry.

For now, let’s talk about non-intuitive feedback instead. Civilization 5’s “warmonger” factor in diplomacy is an example that has resulted in many a player wanting to tear their hair out in frustration. The premise is simple to understand; if you make a habit out of conquest and pillaging, certain other civilizations will be predisposed to dislike you, depending on their own temperament. The execution, however, is not so straightforward. One can defend themselves from an attacking civilization, destroy their armies and take or raze a single city in righteous retribution, and thereafter be reviled as a warmonger by the rest of the world.

A combination of several factors make the warmonger system unintuitive and widely disliked among Civ5 players:

  • It gives no consideration to the initial aggressor. You can become a ‘warmonger’ without ever once declaring war yourself.
  • It gives no consideration to the geopolitical situation. Even your own allies will consider you a warmonger for fighting against a common enemy.
  • Warmonger penalties last an unreasonably long time; civilizations will hold grudges over things that happened millenia ago.
  • It treats the player unfairly. AI civilizations display little evidence of hating AI warmongers as much as they resent the human player.
  • It makes players feel helpless. Rival civilizations can engage in a variety of quasi-hostile activities—spying, spreading their religion, settling on your borders—and the player’s only recourse is often war. With the addition of the draconian warmonger penalties, players must suffer every AI insult to avoid becoming reviled by the rest of the world.

The general opinion is that warmonger penalties are a good idea in principle, but are not applied in an intuitive faction. A player should not be thought of as a warmonger by their allies for fighting on one side of a world war, or for exacting retribution against obviously insulting or hostile acts. This is a good example of poorly implemented feedback, arguably more so than the Europa Universalis scenario I used in the preamble.

There is, of course, another potential game design flaw related to feedback; the absence of it entirely, as Soren Johnson illustrated with the example of the original Civilization’s combat. However, this is quite a bit rarer in modern games than it was in the old days. Games are generally more conscientious of informing the player about the existence of basic gameplay mechanics, even if they often fail to make them intuitive.

I don’t believe it’s necessary to discuss this edge case in detail, and so I’ll see you next week when I talk about randomness.

Leave a Reply

Your email address will not be published.