Game Creation and Balancing Guide
By Savvy-Sauce - Originally written for NSM.
--------------------------------

PART I: PREFACE
1. A Few Caveats
 
PART II: BRAINSTORMING
2. An Idea
3. Construction
4. Tailoring
5. Fine-Tuning
6. Final Thoughts on Game Building
 
PART III: BALANCING
7. The First Check - Claims
8. The Second Check - Lynch Mistakes
9. Win Feasibility
10. Other Balancing Tools
11. Third Parties, Mystery Games and other Wacky Mechanics
12. Final Advice

--------------------------------

PART I: PREFACE

1. A Few Caveats about Balancing

Making a balanced game is incredibly difficult. You have to take an enormous number of variables into consideration, and even then, nobody can be sure whether or not a game is balanced. On top of that, many games feature some form of randomness; the seer might die on night 1, or the vigi could hit three wolves in a row due to dumb luck. Since it's impossible to ensure that the teams in a game have an equal chance of winning, I think the best way to define a balanced game is one for which every team has a feasible chance of winning. That's the requirement I'll be going with for the rest of this guide.

PART II: CONSTRUCTION

2. An Idea

What is the most resilient parasite? Bacteria? A virus? An intestinal worm? An idea. Resilient... highly contagious. Once an idea has taken hold of the brain it's almost impossible to eradicate. An idea that is fully formed - fully understood - that sticks; right in there somewhere.

Every good TWG should start with a good idea. In art, they say you should never touch your pencil/brush to paper/canvas until you know what you're trying to accomplish. This is what separates boys from men, and doodlers from Michaelangelos. The same applies to TWG where you shouldn't write down a single role until you have an idea of what you want your game to be like. Oftentimes this idea is a new gimmick you've come up with: maybe lynches should be done privately, where everyone PMs they're lynch vote to the host? maybe players should have health, where lynches and kills cause different amounts of damage? or maybe you've invented a new role that allows dead players to vote on certain power usages?

It can be anything you want, or a combination of a bunch of different ideas. Just make sure it improves the TWG experience. If you want more help on how to ensure your idea will improve TWG rather than make it boring, check out Liggy's design guide! And if you've already read his guide, feel free to skip 3 (construction and synergy) and 4 (luck avoidance) of this guide!

3. Construction and Synergy

Suppose you go with the idea that lynches should be done through private voting rather than public voting in the thread. These are actually known as ballot box games, but we'll pretend that you came up with the idea all by yourself for this example. The idea leads to a few obvious powers. One that jumps to my mind is a player with the ability to examine another player's day phase vote. Or you could make it more of a researcher, and have him look at another player's vote history. Other ideas might be charismatics (players who have votes which count for more than 1), the wolves knowing all the votes, the wolves being able to manipulate player votes, a role that makes voting public again for one phase, etc.

Pick roles which will enhance your original idea. I would not recommend throwing a seer or guardian in at this point "just because" unless they would somehow help your idea reach its full potential/make the game more fun. If you want to add a seer/guardian those since you think the human team is too weak: just wait. There are other ways of maintaining balance. Let's say you chose these roles for your game:

Ballot Box Game
1. Wolf Boss - Finds out who voted for whom at the end of each day phase.
2. Wolf Shaman - A wolf with seer powers.
3. Wolf Manipulator - A wolf who can change one player's day vote to the player of his choice.

4. Mayor - Vote counts for 3. Can kill himself at any time if the pressures of holding office become to great.
5. Peeker - Can look at a player and find out who they voted for last day phase.
6. Revolutionary - Can make one day phase have public voting. Manipulator and Mayor powers still in effect.

The roles were chosen because a lot of them have synergy with each other--the combined effect of the roles together in one game is greater than the roles individually. If the wolf manipulator targets the Mayor, the wolves could control the lynches for the rest of the game (unless the Mayor wises up and kills himself). And the Peeker can't be certain whether a player is voting against the human consensus in the thread, or if they were manipulated. The Wolf Boss finding out who voted for what can allow them to make intelligent wolfings and form day phase plans, and the wolf shaman can help the manipulator find the mayor. The revolutionary is just a fun guy who will mix the game up for activity. The combination of effects will give the humans a lot to discuss, and the revolutionary will make the game interesting again if it slows down in a later phase.

4. Luck Avoidance

Luck avoidance is a very important part of creating a balanced game when deciding on roles. Take a look at the following game:

1. Master Wolf
2. Wolf

3. Seer-Vigilante - Can seer a player and kill a player every night phase.
4. Human
5. Human
6. Human
7. Human
8. Human

There are some big problems here. Firstly, the humans are extremely dependent on one role to do most of the work, since he has so much more power than every other player. So if the player with that role dies on night 1, the humans will be put at a disadvantage for the rest of the game. At the same time, if a frequently-seered player gets the Wolf role, he won't be able to do anything to stop himself from being lynched. The wolves could also be crippled by an incredibly lucky night 1 kill by the Seer-Vigilante.

This game is very volatile, meaning a few small actions at the beginning of the game can decide its outcome, way before it's close to being over. As a host, you generally want the game to be competitive, and if one team takes a huge blow early on, the players may give up or not try as hard for the rest of the game. Here's a modified version of the above game that is much more stable:

1. Wolf Painter 1 - Single use power. On any night phase, may make the player of his choice appear a different color.
2. Wolf Painter 2 - Single use power. On any night phase, may make the player of his choice appear a different color.

3. Seer - Can seer a player every night phase.
4. Vigilante - Can kill a player every night phase.
5. Back-up - Becomes a seer if the Seer dies first, becomes a vigilante if the Vigilante dies first. Does not know he's the back-up.
6. Human
7. Human
8. Human

By making the wolves ability to appear green a decision rather than innate attribute of one of them, it makes it so that with proper skill, they can avoid being incriminated due to seering. It also makes a red seering less certain, since the wolves could make one of the humans a miller for one night. The separation of the single human power into two different roles also makes the game more stable. If one of them dies, instead of the humans losing all their powers, they would only lose half... except the Back-Up makes it so they wouldn't lose any! Whether or not this particular game is balanced, it provides an illustration of how to prevent luck from ruining your game.

5. Fine-Tuning

Your game is basically done at this point, but you can still make a few small changes. Is it too small for the forum you're planning on hosting it on? Add some humans or traitors. Do you think the humans will have too difficult of time finding the wolves, but don't want to add a special? Try adding cardflipping. Worried that a human alliance could make the game less fun for other human players? Design a third-party role specifically designed to infiltrate an alliance. There's plenty of small tweaks you can make to ensure your game lives up to the concept you envisioned.

PART III: BALANCING

7. The First Check - Claims

Now we move on to determine if the game is balanced. There are two easy checks you can perform to get an idea of whether you're roughly on-target with your game. These are just to fix common mistakes that beginning hosts make until you get a feel for designing games. The first check is simple: what happens if ever special role publicly claims their role in the thread?

In the ballot box game introduced earlier, nothing happens. They'll just be picked off one-by-one in each night phase. But take the following game:

1. Wolf
2. Wolf

3. Angel - Knows he's the Angel. No special powers.
4. Devil - Knows he's the Devil. No special powers.
5. Saint - Knows he's the Saint. No special powers.
6. Priest - Knows he's the Priest. No special powers.
7. Bishop - Knows he's the Bishop. No special powers.
8. Pope - Knows he's the Pope. No special powers.

If all of the special roles claim in this game, the wolves will end up getting screwed over! They'll have to counterclaim a specific role (let's say the first wolf counterclaims angel, and the second counterclaims devil), and suddenly the number of potential wolves has shrunk from 8 to 4. Not only that, but it's usually pretty easy to spot which one is the real angel/devil anyway. This is why new players are dissuaded from creating games packed with special roles, as tempting as it is: not only do special roles usually have incredible powers to use against the wolves, but if the number of claimable roles on the human side becomes too high, the wolves will have almost no chance of winning. Beware of this mistake!

8. The Second Check - Lynch Mistakes

The second check a host should make is how many times the humans can lynch another human and get away with it. Just look at how your game would play out. In a 12 player game with 3 wolves, you know that at the end of Night 1 it will be 3 wolves and 8 humans. At the end of Day 1 it will be 3 wolves and 7 humans. You can continue doing this until the end of Day 3 where it will be 3 wolves and 4 humans. The wolves will kill again in the night, winning the game, meaning that the humans can only mislynch twice if they want to win!

You generally want the number of mislynches to be 2-3 depending on how much help the humans have with lynches.

Although this is generally a good way of determining balance, it can become messy in games with more than 2 factions or in games with multiple sources of death such as brutals and vigilantes. For these cases, you'll just have to give things your best estimate. (I generally just subtract a human from the game if there's a brutal, and subtract two from the game if there's a vigilante.)

9. Win Feasibility

By far the least abstract and most difficult is just to ask yourself is it feasible for each team to win. If the answer is no, you should obviously rework the game, but arriving at that no is difficult.

Try running through the game a few times on paper. It will be tedious, but it's the surefire way of making your game fair. Consider the humans missing their day 1 lynch then hitting the day 2 one, and seeing where both teams stand. Or if you have a vigilante, will him killing an important wolf on n1 doom the wolf team? You don't want a team's chance of winning to sit on a single role, and your game should be flexible enough to pull itself out of a huge lead on one side.

10. Other Balancing Tools

So you found your game to be tilted in favor of the wolves. What can you do? There's quite a few options, and since they're all really intuitive, I'm just going to list them rather than explain them.

• Get rid of a wolf power. (Master -> Regular, Brutal -> Regular)
• Make the wolf powers harder to use (If there's a wolf shaman, add blue humans)
• Add a seer/guardian/other cheap special
• Add cardflipping
• Increase the number of humans
• Swap a wolf for a traitor

And if the humans are the ones who are overpowered...

• Get rid of some humans
• Make the human powers weaker (if there's a seer, add master wolves and millers)
• Increase the power of the wolves (make them all brutal, give them seer powers, whatever)
• Increase the knowledge of the wolves (such as who the millers are, and who some of the humans are)
• Destroy the human alliance (A wolf sniper usually does the trick, but a wolf that goes through a guardian would work fine as well. Or make the guardian fail on blue roles.)
• Add a wolf

11. Third Parties, Mystery Games and other Wacky Mechanics

As the game becomes more complex, so does balancing the game. A game can only be balanced perfectly if you can perfectly predict what's going to happen, and that's usually not the case with mystery games or games with a third party. You have to consider the feasibility of each team beating both of the other teams, and you have to make sure they have the tools to accomplish that. Or if it's a mystery game, you have to consider if any roles will claim, and if an important one does, will it ruin the game? Or a talented player could just lie about his role entirely in a mystery game, dominating in a way nobody could have predicted. 

It's difficult. And it's messy.  Not only do you have to make it possible for the humans and wolves to win, you have to make it reasonable for that third party to win as well, and that’s an area where a lot of hosts stumble.  A quick way to do this is to compare the third parties with the wolves.  Third-party roles are often treated like wolves by humans, so you should balance the game as such. 

If the wolves have a Master Wolf, Revoker Wolf, and Shaman and the third party is charismatic and a vigi, that isn’t really fair.  The third party is about as strong as one wolf, and there’s no way he can take on three!
Some easy ways to make this third party pack more of a punch include:

-Making him immune to wolfing
-Giving him a color that lets him blend in (this should be standard unless you expect this third-party to dominate the game)
-Giving him additional powers (but keep in mind how useful these may be.  Don’t allow him to seer players if he can’t vigi/kill people!)
-Giving him a teammate
-Giving him an easier win condition (perhaps one that makes it so his victory doesn’t prevent the wolves or humans from winning?)
-Combine any of the above

When you think you’ve got a balanced third-party role, double-check that you haven’t thrown the humans’ or wolves’ chances to win out of balance.  If he can joint-win with either of them, that gives them a significantly higher chance of winning—and should be accounted for when balancing.  If he has an incentive to kill wolves instead of humans, then maybe the wolves should have an extra power?  Rinse and repeat until balanced.

12. Final Advice

As with the other guide I wrote, try and compare your game to previous games. Look at a game the wolves won: was it due to a weak performance by the humans, or a design flaw? And consider its impact on your game. Once you have enough experience in TWG, you'll be able to take a look at a game and see if it favors the human team or wolf team.