Last month I looked at the way narrative affects game replayability. This time I'll be looking at how replayability is affected by the game mechanics themselves.

Obviously, the single most important contributor to a game's replayability is its playability in the first place. If a game is badly balanced, if it has a poor user interface, if it seems to be lacking essential features, then it's not going to be much fun to play, much less to play again. But there are specific design considerations that influence a game's re-playability, and those are the ones I'll be talking about here.

Let's start with your basic single-player computer game. Whether a player perceives such a game as replayable depends to some extent on what kind of a player he or she is. Consider our two old friends, the core gamer and the casual gamer. (See my earlier column, "Casual versus Core" for a discussion of these folks.) Since a core gamer's primary motivation is beating the game, as long as the gameplay is interesting and above all challenging, he will continue to play that game repeatedly until he has beaten it, even if the gameplay is very similar every time. The core gamer has no problem with a game like Pac-Man, because even though Pac-Man is a deterministic game that behaves exactly the same way every time you play it, it offers a huge amount of gameplay. Pac-Man contains 256 levels, and very, very few people have ever played them all. The core gamer doesn't mind a game that plays the same way every time, as long as he's got an entertaining challenge to overcome.

This is why arcade games are designed for core gamers, and why they make so much money. Most arcade games provide large numbers of levels and progressively increasing difficulty, and many have deterministic gameplay. The deterministic gameplay allows the core gamer to move swiftly through the early, easy levels, and get up to the harder ones where the real challenge is. Most arcade games are also ultimately unbeatable; they simply get faster and faster until no human being could possibly keep up with them. This means that the object is not actually to beat the game, but to beat your own personal best score, and that's something you can always try for no matter how many times you have played the game. Core gamers give up on arcade games once they become tired of the gameplay or they reach a point beyond which they simply cannot improve, and once a core gamer does beat a game once and for all, he's seldom interested in playing it any more. The pleasure comes from winning, and since he now knows how to beat it, the challenge is gone.

The casual gamer, on the other hand, plays not for the exhilaration of victory, but for the joy of playing the game. It's not enough to simply supply the casual gamer with a tough challenge and let her go at it; she has to be having a good time, and to lure her back again, one thing the casual gamer needs is variety. The game has to be different the next time she plays it.

Sources of Variety

Variety in a game can come from several places:

Varying initial conditions
Most simple board games like chess, checkers and backgammon start with the same initial conditions every time. Both players have the same number of pieces, placed in symmetric positions on the board. But not all games require absolute symmetry. In the board game Stratego, for example, the players start with equal numbers of pieces of equal strength, but they may set them up in their own areas of the board any way they like. This freedom in the initial conditions creates variety for the players.

Initial conditions can also be established randomly; this is of course the basis of most card games. The deck is randomized by shuffling, and then a certain number of cards are dealt out to each player. Bridge and hearts are good examples of card games that depend entirely on varying initial conditions for its gameplay - all the cards are dealt out, and the players play them as they best see fit.

Chance as a part of gameplay
Even if the initial conditions are identical, a game can include random elements as part of the rules of the game. Backgammon and Monopoly are good examples of this. The pieces start in identical positions, but their movement is determined by throwing dice.

Any card game in which you draw cards from a shuffled deck in the course of play (gin rummy and most forms of poker, for example) is using both random initial conditions and randomness during gameplay to create variety.

Non-deterministic opponents
In a game like chess, with identical starting conditions and no random elements, what provides the variety is the opponent's gameplay. This usually (but not always) means that a human being is a more interesting opponent than a computer. Computers tend to be programmed with deterministic, number-crunching algorithms to find the best move, according to some metric for measuring the quality of a given move. With a deterministic algorithm, a computer program will always choose the same move in a given situation. In time, human players can learn to take advantage of this predictability; they also tend to find it rather dull.

Human opponents are more interesting because in addition to having varying strategic and tactical abilities, they differ in the degree to which they're aggressive or defensive, devious or forthright, cautious or risk-takers. And of course, you can talk to them. There's a social aspect of playing against other people that is completely absent when playing against a computer, and that tends to make the game replayable even if nothing else does.

A choice of player roles and strategies
If a player can play a game in several different roles, the game will feel different even if its content is the same. The character classes, races, and alignments in Dungeons & Dragons are a perfect example of this sort of thing. You might play an entire computer game as a lawful good human fighter, then decide to replay it again as a chaotic evil elf magic-user. Although you encounter the same people, creatures, and dangers the second time around, your approach to dealing with them will be significantly different, especially if the designers have constructed obstacles that can be overcome by a variety of methods. (Unfortunately, in far too many role-playing games the only method available is "whack it until it's dead." But at least there are a variety of ways of whacking it.)

Sheer size
You can play an enormous game like Baldur's Gate from beginning to end and still not see every location or undertake every quest, particularly if you concentrate on the main storyline and don't allow yourself to get sidetracked often the first time through. This gives Baldur's Gate considerable replayability. It's just so big that it's worth going back and playing again to follow up on adventuring opportunities that you missed the first time around.

Other Considerations

The most consistently-replayed computer game in the world has got to be Solitaire, the version of Klondike that is included with Microsoft Windows. So what's its appeal?

Some of these characteristics are helpful to us and some of aren't. Item one, for example, isn't much use. Most of us want to design new games, so computerizing existing games from the real world doesn't have a great deal of appeal to us as designers. (It can have a great deal of appeal to those of us who are AI programmers, however. Many existing games make interesting programming challenges - chess is an extremely simple game, with no randomness and no hidden information, but look how much money has been spent on chess programming!)

Item five, too, doesn't help us much. There's not a lot we can do about the fact that Solitaire is free. Most of us want to get paid, so our games have to sell, and that means that there has to be enough content in them for players to justify opening their wallets. Unfortunately, content is expensive to make, and it often lengthens and complicates games. There's an interesting relationship here, one that I think we can learn from: the most replayable games are also the smallest and cheapest to implement.

Items two, three, and four get to the heart of the matter. Summed up in one word each, they are simplicity, shortness, and ease. Trip Hawkins, the founder of Electronic Arts, used to insist that games be "simple, hot, and deep." Simplicity and depth (i.e. subtlety or variety) both contribute to replayability. By "hot", he meant exciting, which is neither here nor there as far as replayability is concerned; it helps if that's the sort of game you like. Solitaire isn't very exciting, but it's still highly replayable.

Personally, I don't think Solitaire is a very interesting game. It's too random. You lose far more than you win and no amount of thinking you can do will change that. Free Cell, which also ships with Windows, is a much better game. It takes a little longer to play, but it offers a mental challenge that Solitaire lacks. Its rules are almost as simple and its user interface is identical. And unlike Solitaire, Free Cell rewards patience and persistence; it isn't that hard to solve to begin with, and in fact all but one of the 32,767 deals of Free Cell can be solved with enough effort. The knowledge that it can be done encourages you to continue to try.

Designing for replayability is the purest test of the game designer. Replayability requires a simple, compelling, addictive challenge and the most natural, frictionless user interface possible. All the big, expensive, fun things that we think game development is about - spectacular graphics, hundreds of unit types, fifteen different camera angles, and voiceover narration by Patrick Stewart - are irrelevant. The game is reduced to its barest essentials: the challenge and the means of overcoming it. If I were trying to design a game for high replayability, I might actually start with cards or dominoes, something I can shuffle around on a tabletop. They wouldn't necessarily end up as cards or dominoes in the game; they could end up as genies or giant worms just as well. Their surface appearance doesn't make much difference as long as the gameplay works.

Conclusion

Replayability is not an absolute necessity for computer games. As my friend Jeff Wofford at Deep Red Games points out, many games offer so much gameplay - forty or fifty hours is not uncommon - that a lot of players don't even finish them the first time through, much less play them again and again. If we've given our customers an enjoyable time for a dollar an hour or so, we're doing pretty well; certainly better than the movies do, even if our customers only play the game once. If I were designing a large game, I probably wouldn't worry about it much.

Still, the question of replayability is one that every designer should ask herself in the initial stages of game design. All players, casual or core, want good value for their money. If the game can be played to its conclusion in a few minutes or hours, then you had better either set the price accordingly, or make sure that it's replayable by design.

______________________________________________________