Quote:

Original post by Edtharan The reason that SPR is such a well used method of balancing a game is that, unless you go too wildly out the standard format, it does guarantee a balanced system.

It isn't a method of balancing a game, unless you keep your mechanics incredibly simple; like the actual rock paper scissors game itself.

As soon as you get into the territory of more complex mechanics such as a real RTS, it doesn't actually balance anything, it just defines a set of vaguely cyclic relationships that suggest (although by no means guarantees) that nothing will be indestructible or useless.

There's no simple mathematical system that will give you a balanced game. Systems like the balance matrix or Wai's point system will give you a good starting point, but they are by no means guaranteed to be perfect. There are a lot of subtle factors that contribute towards a unit's value which none of those systems can take into account. Only thousands of man-hours of tweaking and playtesting can really iron out the imbalances, which is why RTSes ALWAYS ship with imbalances, and it's not until years of feedback from the playerbase that the they get fixed out to the point where they're too small to really care about. Sometimes, not even then.

So to answer the original question, your game WILL be imbalanced. At least to begin with. Any of the systems presented here will give you a reasonable starting point - some better than others - but bear in mind that that is all it is; a starting point. Unit balance will be something that will need to be tweaked, and so you need to design it with that in mind. If you're going to come up with some unit stats, do them in a spreadsheet that can export them in a format that the game can read, and can be easily edited without having to hunt for them in a monolithic design document.

Hi, I am from the perspective where I believe that it is possible to design something that works without ever prototyping it or play testing it. I am not claiming that I know how to do it. But I just want to see whether I could get your attention to think about the question of balance from this direction:

Suppose it is possible to design tactical balance, what is the minimum set of knowledge the designer must have to do it?

When I see this question, I think that the first step is that the designer must have a definition of what "tactical balance" means. What is not defined cannot be designed.

How would you define "tactical balance"? (The original title was play balance, but I think tactical balance is slightly more specific. The original topic could cover balance on aspects more than just tactics.)


Tactics can be defined as:

Tactics Choices of objectives to defeat an enemy based on the assumptions on the enemy course of action

It is implied in the definition that each unit must be capable to serve different objectives during some time in their life of service. In some cases this choice is trivial to satisfy. For example, a soldier can be ordered to attack target A instead of target B. That is a choice of objective in the context of defeating the enemies. In this context, for the player to feel that a tactical choice has been made, the player must have some prediction or knowledge of reactions of the enemy.

Therefore, in terms of game design, it is easier to design a tactical situation in a single player game. In a single player game, the actions of the enemies can follow a fixed script. The player is given options, from which the player forms a course of action to defeat the enemy course of action. The gameplay of a tactical game is the set of mental processes required to:

1) Decipher the enemy course of action
2) Understand the range of available choices of unit objectives
3) Devise a course of action that defeats the enemy's course of action

A tactical game in general would have all three of these gameplay elements, although the engagement value of some element would exhausted sooner than the others. For example, if the enemy has only one course of action, then the mental engagement from element (1) would exhaust after the player there is only one way the enemy would attack. A game that is weak in these mental engagements to a player is not recognized as a tactical game to that player. For instance, Tic-Tac-Toe has all three elements, but all three elements exhaust very soon.

A quick note: In this definition, it is obvious that RPS is not tactical because there is no notion of "course of action". On the other hand, you could see how Chess, Checker, and Go have all three elements. Knowing that "If my Rook could get to A5 I will have a checkmate", is an Element 2 process, while "I will move my Knight to E4, that will prompt my opponent to move its Bishop to B2. Then I will move my Rook to A5" is an Element 3 process.

The above only defines what tactics is. It has not expanded to how to design a game with tactics. But I think it is enough context to define the next term, which is the actual subject:

What is Tactical Balance?

The big things to watch for in balance:

1) Every strategy should have a way to counter it.
2) No strategy/choice a player has should be inferior to another specific strategy in all scenarios. If it is, don't bother putting it in your game.

It's not as important that every option be equally powerful, so long as they're all counterable. A really powerful strategy has the immediate drawback that players will expect it, and basically counters itself in that way.

Quote:

Original post by Wai What is Tactical Balance?

I'll have a go:

Tactical Balance is the non-existence of Dominant Tactics.

A Dominant Tactic is any course of action which offers a significant advantage and cannot be prevented or limited by any reasonable counter action other than itself.

The 'Reasonable' clause is important. It is not enough that a counter exists, it must be reasonably proportional to the tactic it is countering, in terms of difficulty, resource efficiency, obscurity, and risk. The 'other than itself' is also important; if a tactic is it's own best counter, that does not make it OK - it simply makes it a no-brainer to use that tactic as early as possible in every game.

Quote:

Original post by Sandman Quote: Original post by Wai What is Tactical Balance? I'll have a go: Tactical Balance is the non-existence of Dominant Tactics. A Dominant Tactic is any course of action which offers a significant advantage and cannot be prevented or limited by any reasonable counter action other than itself. The 'Reasonable' clause is important. It is not enough that a counter exists, it must be reasonably proportional to the tactic it is countering, in terms of difficulty, resource efficiency, obscurity, and risk. The 'other than itself' is also important; if a tactic is it's own best counter, that does not make it OK - it simply makes it a no-brainer to use that tactic as early as possible in every game.

AKA, "read up on game theory."

As well as having a counter to every tactic, it's equally important to make sure that there is no single option that a player has that is better than all others the player has, in all situations. Even if this tactic is theoretically counterable, if this happens you have a basically broken game.

I don't think it is enough to say that for tactical balance, every course of action must have a counter. That could be illustrated by the following made-up example:

Game: Sequential Rock-Paper-Scissors

In this two-player game, each player takes turn to show a hand sign. Rock beats paper beats scissors beat rock. A player loses when makes more than one change in his turn, takes too long to make a sign, makes an undefined sign, copies the opponent sign two times in a row, or makes a sign that is defeated by the opponent's current symbol in his turn.

Example runs:

A: Rock
B: Paper
A: Scissors
A: Paper -> A loses

A: Rock
B: Paper
A: (Hesitation) -> A loses

A: Rock
B: Paper
A: Spock -> A loses

A: Rock
B: Rock
A: Paper
B: Paper -> B loses

A: Rock
B: Paper
A: Scissors
B: Rock
A: Paper
B: Scissors
A: Rock
B: Rock
A: Scissors -> A loses

You can try playing it with your own hands. You will see that a major component of gameplay is motor-control. In this game, there are only three possible actions: make rock, make paper, or make scissors. The course of action required to win is to have the symbol that beats the opponent's symbol. There are only three possible course of actions that do not led to defeat:

A1) Play the same sign as the opponent this time, and play the sign that beats the opponent's next sign

A2) Play a sign that beats the opponent's current sign, and play the same sign as the opponent's next sign

A3) Play a sign that beats the opponent's current sign, and play the sign that beats the opponent's next sign


There several tactics:

T1) Counter the opponent's symbol during your turn, and hope that the opponent cannot keep up, and messes up on his own

T2) Copy the opponent's sign hoping that the opponent inadvertently makes the inferior sign

T3) Suddenly change the tempo, hoping that the opponent would make two signs in a row

T4) Scream or suddenly shout to make the opponent inadvertently make two signs in a row

When you try this game (with yourself, or with a friend), you would probably agree that this game is Tactically Balanced (even though we having come to a definition). But what makes it Tactically Balanced, is not that each hand sign has a counter sign, but that each player has equal freedom to use all tactics.

The counters of the tactics in terms of tactics:

C1) T1 can be countered by T1, T2, T3, T4.
C2) T2 can be countered by T1, T2, T3, T4.
C3) T3 can be countered by T1, T2, T3, T4.
C4) T4 can be countered by T1, T2, T3, T4.

It is a weird situation where any tactic can counter any tactic. You could also say that the tactics have no counters. But the tactics do not always work.

Is it possible to have Tactical Balance when there is no counter to any tactics? Do the tactics listed above qualify as tactics? How would you describe or analyze this game in your terms?

Quote:

Original post by Wai Is it possible to have Tactical Balance when there is no counter to any tactics? Do the tactics listed above qualify as tactics? How would you describe or analyze this game in your terms?

I think that game fits into my definition perfectly. There are, quite clearly, no tactics that fit my definition of a Dominant Tactic, and therefore, the game would be balanced.

Every tactic can be countered. None of those counters are dramatically more difficult to pull off than the tactic they are countering. Furthermore, none of those tactics are only countered by themselves.

Let's break the game with a rules change. We'll add two new moves.

Spock: Make a sign like a cross between paper and scissors. Spock is far too logical to be defeated by mere rocks, paper, and scissors, and therefore the Spock sign beats all of them.

Fandango: Spock can only be beaten by totally illogical craziness. Fandango loses to all moves except itself (a draw) and Spock, which it defeats. To make the Fandango move, the player must dance the fandango whilst simultaneously yodelling and juggling ice cream.

It's pretty easy to see that this totally breaks the game. Spock is a Dominant Tactic; The first player who plays it gains an unfair advantage which has no reasonable counter; the best a player can manage is to copy, which will ultimately lead to losing as the run of Spocks continue.

What if Fandango was easier to accomplish? Let's make it a simple hand signal again, like the rest of the moves. How does that change things?

It's certainly a lot harder to tell if Spock is imbalanced. At first glance, I think it puts the game back into tactical balance.

The danger with all this pure balancing is, if you get it right, the game becomes boring. There are, after all, no international leagues for RPS. If each tactic is exactly balanced, where is the fun in choosing? You choose one at random, you have a rule, it doesn't matter. Instead, you have to, instead of balancing the tactics, hide which one is dominant, and change it so it never truly becomes dominant. Then it is a matter of balancing how easy it is to judge which tactic is best. Too easy and it is just a unbalanced dominant strategy; too hard and people will just resort to random guessing.

Quote:

Original post by Wai I don't think it is enough to say that for tactical balance, every course of action must have a counter. That could be illustrated by the following made-up example: Game: Sequential Rock-Paper-Scissors

The problem here isn't RPS, it's that you've got a sequential, perfect knowledge game. Those are, by definition, "solvable," although some are extremely difficult to solve like Chess or Go. Though, we're getting pretty close on Chess.

The way to make a sequential perfect-knowledge game interesting is generally to increase the number of elements sufficiently that it becomes difficult to solve.

Apart from that, you either hide some information, or you make the game simultaneous (or real-time, which is effectively the same thing). Randomization can work, too, but that's effectively hiding information (Nature's choice).

Quote:

Original post by thk123 The danger with all this pure balancing is, if you get it right, the game becomes boring. There are, after all, no international leagues for RPS. RPS is just a simple game. It can be interesting to play if you don't play just one round. While a perfectly random strategy can't be taken advantage of, most people are incapable of playing a perfectly random game, and so you can devise winning strategies against other players. If you don't believe me, look up "Roshambot." If each tactic is exactly balanced, where is the fun in choosing? You choose one at random, you have a rule, it doesn't matter. Instead, you have to, instead of balancing the tactics, hide which one is dominant, and change it so it never truly becomes dominant. Then it is a matter of balancing how easy it is to judge which tactic is best. Too easy and it is just a unbalanced dominant strategy; too hard and people will just resort to random guessing.

Dominant doesn't mean that one strategy is better than another overall. It means that one strategy is better than another, no matter what the opponent does.

Let's say we have RPS (not sequential), and decide that winning with rock or paper is one point, and winning with scissors is two points.

Scissors is not a dominant strategy. It still loses to rock, even if that loss isn't worth as many as a loss to scissors. This can make the game more interesting, because now players might expect their opponent to play scissors more often, so they might counter by playing rock more often, etc.

On the other hand, if scissors beats paper AND rock, then paper becomes a dominated strategy. No matter what the opponent does, playing scissors is at least as good as playing paper - scissors beats paper, and they both beat rock. There's no reason to ever play paper any more in the game.

In fact, in this case, scissors is strictly dominant, because there's no reason to ever play anything else.

It's possible to have strategies that dominate others, without being overall dominant strategies. I made a post on this a while ago at http://www.design.wrong.net/2008/05/18/balance-is-not-equality/ (sorry, don't know the bbcode for links).

Re: kyoryu:

There is no problem with the Sequential RPS game. It is a functional game. I was showing that game to show that in S-RPS, there is no real notion of a "counter tactic". The argument was that if a game does not have a system of "counters" then it does not fall into the category of RPS systems and is not compatible to the description that for a game to be tactically balanced, it must have a "counter" for every tactic.

[ Post ] - A previous post on one set of focuses to design a game that has strategies. The set of focuses are: Action, Correlation, Conflict, Consideration and Visibility. The following defines a term.

Action: The designer defines a set of actions assigned to satisfy the strategic aspects of the design.

Correlation: Each action correlates to (but not immediately dictate) a course of action with the objective to defeat the opponent.

Conflict: The players have conflicting goals that cannot be simultaneously satisfied

Consideration: The players are given enough time to consider and analyze the situation. (This is to distinguish the game from a game based on reflex.)

Visibility: The players can observe the actions of the opponent to predict their course of action.

This set does not consider "Tactical Balance", because in terms design, Tactical Balance is an optional objective, it is not a fundamental element of strategy/tactics. (Meaning, So what if I make a game where the undead player always win? It is unbalanced and unequal, but it is still a strategy game. Perhaps the player would define subobjectives and redefine "victory" in that situation: "I survived for 15 days!" "No Way!" Tactical Balance is a popular optional objective, but it is not what makes a tactical game a tactical game.)

So I also agree that tactical balance is not equality. It can exists when things are not "equal", and when things are "equal" it doesn't necessarily exist.


Re: thk123:

What is "this" that you were referring to:

Quote:

The danger with all this pure balancing is, if you get it right, the game becomes boring.

Re: Sandman and others

When I wrote the Sequential RPS rules, I in fact did something to balance the game. I tried to post a minimum set of rules such that the game is not broken. Removing any rule of the described game, would break the game. For example:

1) If you let the player make more than one change during its turn, if the player can take infinite amount of time to make a sign, then in general a player will not lose.

2) If you let the player keep copying the sign of the other player, a player will also not lose. Because copying is much easier than forming a different sign. So if this rule does not exist, this would be a dominant tactic to stay in the game (The dominant tactic is to copy the opponent's move).

When people start playing chess, copying the opponent's move is what they try to do. The logic was that "if I follow your move, then I am at least at the same position as my opponent." This is not actually possible in most games. In Go, it is not possible to copy because the boardsize is odd. In Chess, it is impossible because the ranks are in mirror image (if one copies the movement of opponent Rook, the two Rooks will become face-to-face, and the follower can be captured). It is also impossible to do so in Tic-tac-toe since it also has an odd grid.

It is known that in turn-based games, the player that moves first usually has an advantage. In that context, Tactical Balance must answer how the player that moves second is not at an unrecoverable disadvantage. In Go, this disadvantage is only offset in the scoring (Black moves first and has to score more to win, or that the better player will play White).

In an RTS, it is theoretically possible for both players to commit to the exact same course of action. When that happens, can you say that the game is not tactically balanced? For an inferior player this is obviously the dominant tactic (to copy), because if he does not copy, his chance to win is low. But if he does, the chance becomes 50%. In practice, this is made impossible because the player cannot observe all actions of the opponent and the starting environments are not identical.

When the starting environment, resources, and composition are not identical, we could say that the context leaves the realm of games like Go and Chess. Consider this design:

Game: Zombies vs Human

This is a two-player tactical game. The actual rules of the game are not important at this point. The important details for now are:

1) Both players can see all actions that the other player does
2) The players cannot copy the actions of the other player because the two sides have different action sets
3) At any moment in the game, a player could pause the game to give orders
4) Statistically, no matter what side is played the player has 50% chance to win

(I added rule (3) because I want to disregard the argument that the player that clicks faster wins.)

While (4) asserts that the game is balanced, it says nothing about whether the game is Tactically Balanced, because as far as details go, the game could involve just having the zombie leader and the human leader meet in the center and play RPS to settle the conflict. Consider the following details:

Detail1) Suppose there is a dominant tactics for both sides, and if both players use their dominant tactics, the statistical result of (4) is achieved. The dominant tactic for the Zombie player is to exert minimum control so that the zombies would spread and convert more human on their own. The dominant tactic for the human player is to stick together so that they can effectively kill the zombies.

Detail2) To beat the human player, the Zombie player needs to observe and predict where the human player is going to defend, and weaken those places accordingly before the human defence is too strong. The Course of Action that the zombie player observes involves observing where the human player intend to fight and what traps and attacks the human player plans to use. To employ this tactics, the Zombie player must at times concentrate attack using the zombies, although that would make the zombie population grow slower.

Detail3) To beat the zombie player, the Human player must hide his plan sufficiently so that the Zombie player cannot predict the attacks and the locations where the zombies will be decimated. To do so, the human player must at times spread out to gather items and resources through out the map. This move would make the humans vulnerable to attacks.

Is this game Tactically Balanced? Each side has only one dominant tactic, although the tactic required to counter it is not itself. It is not itself simply because the two sides are of a different race. Is there enough information to decide whether this game is tactically balanced?

[Edited by - Wai on April 19, 2009 8:35:20 PM]