Action games for the designer stuck in the 21st century

Scoring

Ever since Doom, there has been a tendency in games not to keep a proper score. This, in my view is a mistake. Keeping a score changes the dynamic of a game; you can complete a level, or you can complete it well. A score rewards players who achieve the latter.

Really high scores give the player a better sense of achievement, so you should give points for as many achievements as possible. If an average player does not score at least in the tens or hundreds of thousands, you are probably not giving out enough points.

You should never, ever subtract points for any reason;

• Subtracting points is unfair; the player earned those points, and if they had lost the game a moment earlier, they would have scored higher.
• Subtracting points leads to lower scores, and lower scores reduce the 'just one more go' factor.

Whenever you are considering subtracting points for a certain event, consider instead subtracting a life or giving a bonus for avoiding that event.

If the score always ends in some arbitrary number of zeroes, you are not being imaginative enough. If the player always scores in multiples of 1000, then a score of 237000 might as well be presented as 237. Players are not stupid and such scores are much less satisfying. (The same applies to scores which are always a multiple of any other arbitrary value).

The best way to give satisfyingly large scores to players without falling into the above trap is to make the score for any given achievement variable depending on factors that measure the player's performance. For example, if your game is fast paced and you want to reward quick reactions, you could score the destruction of an enemy based on a function of defeating that enemy and the time the player took to do so. In a game with a really good scoring system, it should be possible to score in the millions and yet still beat another player's high score by one point.

The range of scores should be as wide as possible, so as to differentiate maximally between player abilities. An absolute beginner should score in the hundreds or low thousands, while a seasoned expert should score in the high millions. If your scoring system gives a score in the hundreds of thousands for most players, then it is not giving as much information as it should.

The most obvious way to improve the range of your scoring system is to give exponentially higher scores even for the same action as the game progresses. The problem with this approach is that a player's score in earlier levels becomes exponentially irrelevant later on. A popular approach that avoids this problem is to use a multiplier system, which rewards good performance (for example; destroying multiple foes in quick succession) by multiplying any points the player receives by an increasing factor either for a duration of time or until the good performance ends.

Reward the player visually

If the player does well, let them know with spectacular graphic effects. Don't waste your best effects for the player's death; save them for those of the enemy hordes.

Up the kinetic ante! If your game is fast paced but takes place in a small arena, throw the arena around the screen relative to the players movement or actions. This makes the simplest of games surprisingly exciting; see Tempest 2000 and Quadnet for examples.

Don't try to mimic reality

Reality is boring. The more time you spend trying to be real the less you are spending having fun.

Design the game first. If you can think of a realistic setting where your design might work, then by all means go ahead and use it, but never let reality drive your design. In Quake 3, falling a huge distance does you hardly any damage because the game would be boring otherwise.

Some of the best games are totally abstract. If your idea works best in an abstract setting then it is probably far more original than anything that has been released in the last ten years. I know of very few really good games with a physics engine.

Breaking up the action

There are two things that should never, ever appear in an action game; cutscenes, and anything that in any way resembles a cutscene.

If you absolutely must insert a cutscene, the worst possible thing you can do is give the player control of a character with absolutely nothing constructive to do. If I wanted to run around aimlessly shooting the walls while someone I don't know rambles about something I don't care about for half an hour, I'd borrow my housemate's Nerf gun and attend one of my lectures for once[1]. The first 20 minutes of Half Life 2 is one giant cutscene, and this is why the game is boring.

If you feel you need to give the player a breather, a much better way is to insert bonus levels. Bonus levels are much easier than the rest of the game, but they still allow the player to control the action, and they can be very rewarding for a skillful player. However, bonus levels get boring quickly and should be used sparingly.

Dogma

Be creative. Dogmatic adherence to these rules will make your game tired and predictable; surprise the player by breaking them occasionally.

---

[1] This would still be a lot more fun than an 'interactive' cutscene.