Thinking Virtually

#64: Designing Strategy: Victory & Balance

by Shannon Appelcline
Mar 07,2003

Last week I talked about the various phases in a strategic game, from when you lay out the board and arrange the pieces, to when you sweep everything back into the bag. And, in doing so, I purposefully skipped over two phases toward the end of a game: the aptly named end game and the victory phase.

At the time I said that they were a topic in themselves, which is true. Even more important is the fact that the whole idea of victory in a game expands far, far beyond the end game. You have to think about how your game's end will play out from the moment the first token is placed upon the board.

As I progress toward the final calculation of victory, I'm going to talk about some meta balance issues which affect every aspect of play and also some specific aspects of balance related to player involvement and player interference. Then we'll get to the end game and see what types of problems you can discover in the last quarter or so of your game.

Types of Meta Balance Problems

Clearly, no game is ever actually balanced. Some player will be more intelligent than the rest, or another will be able to play more strategically. One player might not have gotten enough sleep the night before, and another might have an innate advantage just because he happens to think like the designer.

Except in the most dystopian science-fiction stories, none of that's your problem. Instead, when thinking about balance, it should be your goal to make sure that no element in the game itself is innately unbalanced.

Below I've compiled a list of some of the most common balance problems, which I call meta balance issues because they can affect the entire play of your game, not just the end game and the victory phases:

Starting Imbalance: Is some player innately favored because of his starting position in the game. Is some power marker given to players at start better than the rest? Is some environmental position favored? How about some limited token? (In Galactic Emperor: Hegemony, one player collected stats saying that the player down in the southeast corner of the map never won.)

Play Order Imbalance: Related, is the player who gets to go first innately advantaged (or disadvantaged)? Or, less commonly, is there a specific sweet position (say, the third player to play) who tends to always do the best? Or the worst?

Random Imbalance: Is there some random element in the game (usually a card drawn from some sort of deck) that is innately "better" than any other? Is there a "win-the-game" card of any sort? Does a particularly unlikely die roll give a player a notable lead?

Power Imbalance: Does a player tend to do better in the game the better he's doing? Is the winner in a position where he innately increases his lead step by step?

Meta Balance Strategies

In all honesty, you will have imbalances even in your game design. The idea isn't necessarily to eliminate them all, but rather to account for them, so that your game remains fair. The following strategies seem to be some general ways to offset meta balance issues:

1. State Your Imbalance. If in some way you make it very clear that someone is in a more powerful state, the other players in the game will act as your balancing mechanism by trying to bring the leader down (though also see my discussion of player interference, below). The Settlers of Nurnberg, for example, has a victory track that records what everyone's built; players can thus see who has the "power imbalance". Some games actually purposefully accentuate this imbalance to really make players mad and encourage them to overthrow the winner. The Great Dalmutti is one of the best examples of this, because it gives the winning player the most comfortable seat in the house and the losing player the least comfortable one at any time. (Applicable to all four meta balance issues, though most widely applied to power imbalance, which might have originated from other balance issues.)
2. Give Your Advantage Disadvantages. Alternatively (or in addition) you can make sure that anything which has an innate advantage, has a disadvantage too. If a power marker is really powerful, give it something that hurts the player. For example, in The Settlers of Catan the player who places his starting piece first takes advantage of a clear starting imbalance, because he'll place his settlement on the best space on the board; however to offbalance this he places his last settlement last of everyone. (Most applicable to starting and playing order imbalances, though somewhat applicable for the rest.)
3. Offer Multiple Paths to Victory. One of the best ways to control imbalance is to make it not matter, and you can do that by allowing multiple paths to victory, so that whatever unlikely event occurs, a player can try and build on it. Poker is a fairly clear example of this: if you get multiple cards with the same number, you build toward pairs, 3s, or 4s of a kind; and if you get multiple cards with the same color, you build toward flushes. Carcassonne also shows this type of design: depending on whether you draw a good city, road, or field piece, you'll make different decisions about how to play. (Most applicable to random imbalance.)
4. Explicitly Help Players Who Are Behind. Some games perfer to take a more explicit tact and have elements in the game which help out players who are behind and/or hurt those who are ahead. For example in The Starfarers of Catan there are some power markers you can win from the aliens which allow you to take spaceship pieces from the player with the most victory points. Other games allow players without money (or something else important) to "beg" to earn that piece from the bank--or from the player with the most of it. (Almost entirely applicable to power imbalance.)

You can actually use every one of the core strategy systems which I've discussed in previous articles to help create balance in your game.

For example, take the issues of timing which I discussed last week. You can use turn timing to offset power imbalance by always ordering turns in such a manner that the players that are the furthest behind get to take the most advantageous position in the turn order (usually that means going first, but it also could mean going last in a game where it's advantageous to see a player's move before you go).

To take another tact, consider the random factor of three weeks ago. Wizwar has a "chaos" card which can be played to force all the players to turn in their cards, and then randomly redistributes them; clearly this will only be used by a player who's behind to try and get better cards (from the "winners"). Other games might randomly give out markers or tokens which will be benficial to people who are behind, because they don't have many of them, but which will be irrelevent to "winning" players who are already stocked up.

Ultimately, you need to think about balance because you're creating a competitive game, and people need to feel that, as they move toward victory, they remain on a level playing field.

Balance & Player Involvement

Inevitably the meta balance issues will closely interelate with game players, and how they react to balance. You can have a serious problem with player involvement, wherein power imbalances can lead to bored players. However, players can also be the savior to your game's balance if you build in rules for player interference (though that causes problems of its own).

We'll talk about player involvement first.

You can have bored players, and thus lowered player involvement, if you either knock players out of the game or allow their position to degrade to such a point that they can't possibly win, but must keep playing. Which situation is worse is somewhat debatable.

If you're using some of the meta balancing techniques I've already described, you'll hopefully be pushing players toward the middle. Some other solutions that can help solve the problem of player (un)involvement are:

1. Carefully Think About if Players Should Be Eliminated. I can't just say "never eliminate players", because it's actually OK in some types of gameplay, and adds nice aspects of risk and reward. Generally, it's acceptable to eliminate someone when you're already in the end game phase, or if they can come back into the game as a new (if disadvantaged) player. Beyond that, elimination is usually acceptable if you're running a game that's either geographically remote (PBMs, PBeMs, or web games), or chronologically lengthy (many days), as in either case the game isn't centered around a single social occasion where you're trying to keep everyone involved. And, the nice thing about allowing elimination in all these cases is that you have to worry less about players getting bored, because if they're losing, they'll get knocked out. However, for more standard strategy games, which tend to take between a couple of hours and a day, and are done around a tabletop, you should probably avoid elimination if you can. I think Wizwar is a good example of a game that ignores this rule to its deficit; any character can be killed at any time, and after that there's no real chance for participation, thus players can be left twiddling their thumbs for an hour until a new game begins.
2. Hide Some Victory Information. Even if a player can't win a game, he'll still enjoy himself if he doesn't know that. Thus, it's often useful to hide at least some of the victory point information. Keep enough visible that players can have a general sense of what's going on, but not enough for a player to decisively know that he can't win. Tigris & Euphrates is a good example of this, because you get to see players accrue their victory points, but the physical representations of the points are hiden behind a screen. Carcassonne is an example of a game that obscures Victory Points by pure complexity; the scoring period at the end of the game takes 5 or 10 minutes to compute, and thus no one can really figure it out during gameplay--but at the same time some percentage of the points, earned during the main game, are marked on a score board and thus players can use that as a basis for "who's winning". (Though, as a caveat, it's possible to overplay this tactic; if you make victory information entirely opaque, to the extent where a player really doesn't know how he's doing at all, you risk turning him off entirely, a mistake I think Skotos made with its early Galactic Emperor: Succession game, where the activity-to-victory correlation was very hard to fathom.)
3. Make Victory Points Transient. If there is some way for victory points to exchange hands, even if it's unlikely, this will keep all players playing on, feeling that, until the last die is cast, there's always a way to gain back the lead. You clearly want to allow players who are ahead to keep the advantage, but even having that small chance is enough, as I'll discuss in a second. Carcassonne takes advantage of this idea too, by allowing fields, cities, and roads to be controlled by the player with the most tokens, and thus giving them the opportunity to change hands up to the end.
4. Control Risk-to-Reward Ratios. It's a general truism that the greater the risk, the greater the reward. You can use this to your advantage to help players stay in the game, especially when combined with the transient victory points that I just suggested. If a player is way behind, you always give them some opportunity to catch up, even if it's really, really unlikely. You should probably exponentially increase your risk in relation to your reward, so that the player who's ahead has a strong advantage, but you can still allow for some way to catch up. Formula De is a pretty good example of a game that allows "hail marys" like this. When you're behind there are usually small chances for you to catch up; for example you might need to roll exactly a "21" on a die numbered 21-30. If you succeed, you catch up with the front drivers, but if you fail your car blows up in a mass of expended tires and brakes. It's a very high risk and the payoff chance is low, but the reward is getting back into the game.

Balance & Player Interference

When considering player involvement you have to balance your game carefully to keep players enjoying themselves. However, there's a flipside: you can also allow players to help keep the game balanced through player interference.

Player interference is, quite simply, a way for one player to have an adverse affect upon another. Without player interference, you might as well have a bunch of people playing simultaneous games of solitaire, which can be very frustrating, because everyone can see one player taking a lead, and no one can do anything about it. With player interference any group of serious players will constantly be ganging up on the winner in an attempt to make sure that no one's dramatically ahead as you move into the end game.

You can allow for player interference in a number of different ways:

1. Allow Direct Avoidance. Just by allowing players to interact in some mutually beneficial way you'll innately set up a method whereby a "winner" can be punished. The Settlers of Catan shows a good example of this: players regularly trade each other for required resources, and as a game moves toward conclusion the leader is left out of these trades with increasing frequency.
2. Allow Indirect Interference. If you build a game with limited resources of any sort, then there will be indirect intereference in your game: you can try and take a resource--be it a space on the board or a card--that you know the "winner" wants. This is typically the greatest level of interference allowed in "family" games, which tend to be built with lower competition levels.
3. Allow Uncertain Interference. This is one of my favorite types of interference, where you can interfere with something that you think hurts another player, but aren't certain. Citadels, for example, lets you steal from a certain character or kill a certain character, but you don't know at any time which player is playing which character. Titan: The Arena allows you to attack arena monsters, but you rarely know which monster each player has their "secret bet" on, and thus your interference is never a sure thing. This is somewhat related to the idea I already mentioned concerning hiding victory information, and goes to a topic of "information" that I haven't really discussed much to date.
4. Allow Direct Interference. At the height of interference is the ability to directly attack another player, usually by destroying their tokens or taking their environmental territory.

Ultimately the question of interference level will be somewhat related to what type of game you're designing--but if you want a fun, balanced strategy game, I'd suggest always going up to at least level 2.

Interference & The King Maker Problem

However, introducing player interference doesn't come without problems of its own. (Isn't that always the case?) The biggest of these is that you introduce the possibility of king makers.

The king maker is a losing player who gets to decide who wins, even if he can never win himself--thanks to levels of player interference being sufficiently high in your game. This is a more common problem in games which remove locality from some aspect of the game. If you can trade with everyone in the game or attack everyone in the game or take some other important activity regardless of geography, then you can make arbitrary decisions about who to help and who to hurt.

A few ways to eliminate this problem include:

1. Control Locality so that Players Can Only Affect their Enemies. This basically subsumes the king maker problem into part of the risk assessment of interfering with another player. If a player tries to eliminate an opponent, he innately knows that the opponent can take a "last strike" against him, which will affect him adversely. Wizwar is a good example of this. You can only attack players whose wizard you can see, and thus you're generally only going to be hurting the person who's hurting you.
2. Control Beneficial Player Interactions. Consider carefully before you let players freely exchange benefits. Instead you probably want to constrain them by something within your game system. Wizwar is again a good example of this idea. You can trade cards, but only with a player whose wizard you can see. If one player is killing you, you're much less likely to be able to dump cards to another player.
3. Decrease Player Power as His Victory Decreases. If a player is less able to do beneficial (or destructive) things as his place in the game decreases, then by the time he realizes he can't do anything to win he won't be able to affect the outcome one way or another. Galactic Emperor: Hegemony in general shows this idea. As a player slowly loses ground in the game, his number of worlds decreases, and thus he has less income to increase his wealth and less factories to build fleets. Thus, by the time he's lost he can make a small final strike (against a localized opponent) but not much more.
4. Make Kingmaking Explicit. Finally, as with many game design "problems", you can integrate the issue into your game. If, for example, a kingmaker gets some bonus which might help him in future rounds of play, then kingmaking becomes a strategic decision where you're trying to throw your heft behind the winning player. A game could easily be imagined where it's the kingmaker who wins, rather than the king himself. (In fact Avalon Hill's Dune game incorporates this idea. The Bene Gesserit player gets to prdict a winner and a turn on which they'll win; if that player wins on that turn, the Bene Gesserit wins instead.)

Types of End Game Problems

All of the topics that I've covered thus far--meta balance, player involvement, player interference, and kingmaking--can affect the entirity of your gameplay. However, you'll eventually arrive at the game phase that I labeled "end game" last week. This is when it becomes increasingly obvious that someone is going to win, and players begin to make their last ditch struggles to emerge at the top.

There are several new problems that can emerge here if you're not careful:

Setback Problems: Most games do have some player interference built into them, as I described in the last section. Although generally good, these can cause problems here, in the end game, if they're not well designed.

Limited Setbacks represent one side of the spectrum. In this situation there are a limited number of ways to stop a winner, usually represented by stop-the-winner cards drawn from a deck during the play of the game. The problem with this setup is that after all the setbacks are used up, the next person to close in on victory is the one to win. Generally, this isn't very fair because it means the 3rd best player (or the 4th or 5th or whatever) is the one that actually wins, rather than the best strategist.

Unlimited Setbacks go to the other side of the spectrum. There are so many ways to setback a winner that no one ever wins, and play continues on pretty much forever, or rather until every one is so sick and tired of the game that they finally let someone win out of desperation.

This all seems to be a bit of catch-22, but there is actually a middle ground where you can allow for unlimited setbacks, but control their power so that they can't elongate the game indefinitely.

Here's some ways to avoid setback problems, based on controlling their power:

1. Offer More Positive than Negative Possibilities. Generally if a player moves forward 3 steps for every 2 he's knocked back you'll be OK, because the game will slowly move to conclusion. However, you may not be able to calculate this as a simple ratio, if player interference isn't controlled by locality. For example, if your game allows for up to 4 players, and every player can adversely affect every other player, than positive cards must be more than 3x as common (or 3x as useful) as negative cards, because a "winner" could potentially have 3 bad cards played on him every turn.
2. Allow a Player to Grow Beyond Setbacks. This is perhaps a different way of looking at the same solution. You can allow setbacks to indefinitely setback a player as longer as he can continue to grow his game position to a point where that's no longer possible. Fist of Dragonstones is a game that supports this idea. There are 3 core ways you can gain victory points: by turning in 4 different dragonstones, by turning in 4 identical dragonstones, or by turning in 3 dragonstones, one of each color. As a player reaches a point where one of those routes could be used to achieve victory, other players can make sure that victory isn't available to him (by stealing the character that would allow the victory condition), but as the player matures his position he can reach victory via multiple paths, and eventually the other players can't stop them all.

And here's some ways to avoid setback problems by fiddling with when the game ends:

1. Put a Timer in Your Game. Totally orthagonally, you can worry a bit less about unlimited setbacks by putting a timer in your game which will eventually signal game end, even if no one has reached the designated victory conditions. Many of the Settlers of Catan games do this. In Settlers of Canaan the game ends when the wall at Jerusalem is completed, and in Settlers of Alexander the game ends when the title character reaches his final battle. The recent Risk: Lord of the Rings offers a similar mechanic where the game ends when the Fellowship reaches the Cracks of Doom.
2. Hide the Victory Condition. If players don't know quite when to offer a setback, setbacks implicitly become less powerful. Some games do this by randomizing when the game ends. Starweb is an example of this; the game ends at 10,000 victory points, plus or minus a fairly wide percentage. I've considered a similar mechanism in Galactic Emperor: Hegemony, to change the duration of the game from 21 days to 19-23. Other games simple give different players different victory conditions, and don't reveal who's trying for what. Mystic War is an example of this: you draw a victory point condition at the start of the game, and thus could be playing for 40 mystic power, 40 gold, 40 population, or else a balance of the three. This technique can also serve some purpose in alleviating some of the boredom and kingmaker problems that I earlier discussed, because of the uncertainty introduced.

Chaos Problems: A less common but real end game problem is that of chaos; even though a winner might be "evident" leading into the endgame, random and arbitrary factors could be so large in your end game that this can be totally changed before the game is resolved (and tends to be in true chaotic games).

One of the "setback" problems I discussed actually falls into this category too. When you have limited setbacks, you're really creating a chaotic situation where the winner is fairly random.

Some general solutions to chaos problems include:

1. Increase the Number of Individual Random Events. Randomness will tend toward an average, expected value. Thus you can paradoxically decrease the amount of randomness in your game by increasing the number of random events. (Or, to use mathematical terminology, you decrease your variance through increased sampling.) Risk is a good example of this. Since you roll dice for every couple of armies, you could roll dozens of times just in one attack; your variance will average toward norm in any individual assault, and definitely in the game overall.
2. Control Your Power Curve. Some games have a very steep power curve at end game, where chaos is created by the fact that each random result is a lot more powerful/important than the one before it. Sadly, Risk, at least in its older variants, offers an example that flies in the face of this problem, rather than solving it. Every army that you redeemed for a collection of cards was larger than the one before, and thus you created a very chaotic end-game situation where armies swept back and forth across the board in mass, and the ultimate winner wasn't necessarily related to who was ahead just a turn or two before. (And this is really another way of repeating what I said last week, when I talked about making early parts of your game important.)

I won't go on much more about chaotic end=games, because I think I already covered possible solutions a couple of weeks ago when I discussed the random factor.

Simply, be aware that the problem exists.

Conclusion

Game balance and end game issues are both pretty big topics (as the preceding 4500 words no doubt demonstrate). I've offered a lot of solutions this week, and you'll be able to see which work for you on a case by case basis. If everything else fails, I think the biggest points are: to make game imbalances obvious; to give players the opportunity to interfere to correct those game balances; but to ensure that interference doesn't result in additional problems, from king making to indefinite extension.

Before I close up I should thank RPGnetters Cerebus, who helped get me thinking about some ideas of player-induced game balance in a response to a previous column, and Mark Green, who reminded me about the king maker issue in a response to a review of Tigris & Euphrates Thanks also to Skotosians Quigg and Christopher Allen, who offered some valuable thoughts on the first draft of this article.

For my next discussion of strategy, in a week or two, I play to cover the topic of information, and maybe some other subtleties that I've missed along the way, if there's space.

And after that it'll be a few different codas and conclusions before I bring this series to an end (whew).