Apr 6, 2014

Designers do not Play Dice with the Rules

Last week I made a case for randomness in Roleplaying games, even for those with a focus on narratives and tactics, which are usually not random. To summarize, my points were:

Randomness adds an element of surprise. You never know for sure how well each individual roll of the die will go, it might spectacularly tip the odds in your favor or it might fail to achieve anything. When you cannot anticipate all that is going to happen then things are bound to be more exciting.

Randomness is fair because dice don't play favorites. In a conflict between characters of perfectly equal strength the winner will be the one who can take the most advantage of high rolls and mitigates better the bad rolls.

Randomness makes things less repetitive. If every battle against every enemy went exactly the same way, you would get bored of them pretty quickly. Your guns might get a lucky critical hit one moment and blow up in your face the next.

If we don't apply these traits correctly then our game becomes exactly the opposite. If you're not careful, randomness will transform your game into nonsense.

Nonsense can be fun though. Just make sure you do want nonsense.

Slice of the Dice

A good RPG knows just how random it wants to be and designs its rules around that. Some games are highly random while others save the randomness for very few specific instances. They are both opposite ends of a spectrum. I am going to gloss over Roleplaying Games that have no random elements to them and those that are entirely and one hundred percent random though. There are such games, but they are too extreme to compare them to BCG.

Deterministic games leave very little to chance, usually they are built around the idea of predictable outcomes and dice rolls all have a bell curve probability to them. You are usually in full control of your character and can reasonably tell in advance whether you'll succeed at something before you do it or not. These games tend to be better when their focus is on a strong coherent narrative and with an eye on game balance.

Chaotic games instead leave a lot to chance, seeking to create a strong play experience full of surprises. You usually have very little control over what happens since your own character is randomly generated. If Deterministic games are a walk through the park, then Chaotic games are a rollercoaster. These games are better when used to roleplay the type of fiction where the PCs are lowlives in a vast and uncaring universe where everything can and will go wrong.

Most games are, of course, somewhere between both of them. Battle Century G is closer to the Deterministic side of the scale. It solves conflict by using ten-sided plus a number between 1 and 10 to try and get a result between (most often) 11 and 21. You can add more modifiers on top to make sure your result is high, and sometimes you want to roll low while other times you want to roll even  but the gist of it is that the dice are somewhere between one third and one half of what makes your result - the rest is up to your Character.

Because it is a game with an emphasis on big flashy heroics and a smooth narrative, this works out just fine. If it were a game about insignificant nobodies who could die at the drop of a hat with no rhyme or reason because life is cheap, then it would need vastly different rules for conflict resolution.

This method would represent the idea perfectly.

Running the Numbers

I went with the d10, but I could have used something different.

I could have used a d6 instead of a d10, with lower Attributes (from 1 to 5) and Advantages becoming being transformable to a +1 instead of a +2 to compensate. That way the game would have a more coherent flow of narrative to it, leaving less things up to chance. I didn't, though, because the game is already Deterministic enough as is. You have Genre Powers to turn a failed roll into a success, several abilities that grant you Advantages to let you keep the better roll or add flat bonuses to the result, and other fancier stuff. If I were to make the game any more Deterministic it would be too easy to "solve" it like you would a math exercise.

Likewise I could have used a d20 and raised the Difficulty Number scales to range between 10 and 30 instead of 1 and 20, that would have made the game swingier and more exciting. But I didn't, because it is much harder to work with it on a mechanical level. Lower Attributes (between 1 and 10) would be worthless when compared to your die result, while high Attributes (between 10 and 20) would automatically succeed at many tasks. You could combine both of them (between 1 and 20) to get the best of both worlds, but you also get the worst of both worlds in that sometimes the dice decide everything and sometimes they are meaningless.

Instead of thinking about what the game would be like with d6's and d20's, let's think bigger. Let's think d100. For those who don't know, a d100 is two d10s but using one as decimals and one as units. This is often used as a "Percentile" system, where you want to roll under a specific number (say 50) and your odds of success that are an exact percentage (in this case, 50%). Percentile dice are very good for modeling a complex world with complex characters, because there is just so much range between 1 and 100 you can add a lot of detail to things that way.

The problem is that it is too complex and detailed to play smoothly. Making every 1% of that 100% matter is, to put it lightly, difficult. If you are not going to make every 1% matter and you'll just stick to making things come in chunks of 5% or 10%, then you might want to use a d20 or d100 instead. Percentile is also very, very random. It has more variance than the d20, though it fortunately has flat odds of success so it doesn't have the problem where your stats are either too high or too low to be comparable to the dice roll. It does work much better in Chaotic games though, especially those with tables of random results. I like percentile but it is difficult to design with it.

I could also use "dicepools", which is to say, have every Test involve lots of dice. Said dice can be added up to make a greater result, adding multiple d6s together (often between 2 and 4) to create a bell curve effect. You can roll very low or very high, but most of the time you will roll average. Much like with just having 1d6, I don't want the game to be that predictable by default.

There are also systems featuring "Exploding" dice. A die that explodes is one where getting the highest possible result means you get to roll again and add the new result on top of the original number. You roll a 10, and that lets you roll another d10, and getting another 10 lets you roll another one again. The name comes from, presumably, the fact that your totally average roll can randomly explode into a ridiculously high result. This is very exciting, as you can imagine, because it means you can poke a man with a stick and watch him explode in hilarious fashion, but it is also very difficult to balance because, well, a man can explode from being poked at. It is fun to play with but a little too random for my taste.

There is also one other method to solve things in a random way: Cards. You could have everyone draw cards at the beginning of an Operation and play them instead of rolling dice, using the appropriate number as the result. It adds a whole new element of strategy to it, because you want to save your best cards for your more important plays but at the same time you can't just play 1's early on because you'll just get killed. There could be ways to let you draw more cards and replenish your hand during your Turn too. The GM too would get to play like this, having more cards than the other Players and likely saving the good ones for the Boss while letting the Grunts die. It would make everyone perform much better when it counts, and I'd love to give designing a game like this a try sometime. The problem is that I do most of my gaming online, and decks of cards are harder to simulate with a chatroom bot than dice rolls.

There are also the d8 and d12 but I like to pretend they don't exist.

Seven Years of Bad Luck

No type of conflict resolution is inherently better than others. As long as you are aware of how each method works and what type of Roleplaying Games it is best suited for, you will do good. I know I've skipped over a few, but these are the ones I considered briefly for use with BCG. There are a few things that you should not do with randomness though, they are the "Taboos" of luck-based mechanics in a way, and anyone who breaks them will suffer frustrating gameplay and annoyed Players.

Randomness should always have an upside. The fun part from rolling dice comes from the possibility of getting a ten, and this is particularly true with exploding dicepools, making it akin to how playing the lottery means you could maybe win the big prize. What you should not do though, is to have an equivalent where getting a 1 means you automatically fail with dicepools. An in-built rule in the game that penalizes the PC for rolling badly, turns the excitement from rolling dice into tension and suddenly you don't want to gamble at all, which in this case means you don't want to play the game.

If you want to introduce tension to the rolls, you should make it a thing that happens infrequently like with Miracles or Overheating Weapons. If you want to introduce a random element that penalizes the Player (like say, a table to determine which mental illness you get from enduring a horrific experience) it should only have negative outcomes. That way you can still get "good" results because you're not getting the absolute worst ones you did not want, without being frustrated that you're still getting a bad thing.

Do not manipulate other Players' luck. Everyone has ways to manipulate their own randomness, and the game more or less revolves around hitting enemies for lots of damage and avoiding getting hit for lots of damage. To that end you can do lots of things to increase your odds of success and decrease those of your opponent. But the one thing that is barren from the land of mechanics is outright deciding results for your opponent.

If dice rolls are fun because you can get a ten, being told in advance that you can only get a five (or lower) in advance is really, really annoying. An ability that makes you tougher to hit by buffing your Defenses is powerful but feels right. Deciding someone else's dice roll result for them could be balanced until the intellectual half of our brains is fine with it, but it will never stop being annoying to the emotional half that feels robbed of the chance to make a cool move.

Frontload randomness and avoid making it the deciding factor. Random complications are fun when they happen early on and you can react to them. Randomness is part of the journey, but cannot be the destination. One dice roll should not be what resolves the climax of the game, that's why all Players have so many abilities they can use to power through bad rolls with bonuses or to fix a bad result with a reroll. If your character dies without you getting a chance to respond to it in any way then you feel like the dice robbed you of a character, but if you get a chance to respond to it and make a mistake you feel like you learned something.

This is why random character creation is so popular: If you get a character who is better than average then that's fantastic, but if you get a character who is below average then that's your chance to prove you can make it work because you're just that good. Even when it does not involve the rules and you're only randomly determining fluff traits, the challenge of making a strange combination of traits work is a fun exercise in itself.

Break these Taboos at your own risk.

Ultimately randomness is a double edged sword that makes things more fun as often as it makes them less fun. Battle Century G wants to have its cake and to eat it, because it has a fair degree of randomness to it but also gives you plenty of ways to control the results. If you want to write your own game then deciding how random it is and to what end it will use randomness is one of the most fundamental aspects. 

Until next week, may the dice be with you.


  1. Exploding Dice actually add very little to the average range in the long run.
    Exploding at -1 (Hackmaster's Penetration system) for example only increases the die average by 0.5. So 1d4p averages 3 instead of 2.5

    One alternative could be to scale the dice. If 1d10 feels too random at low tension on PL1 mechs, but doesn't feel as though it affects things enough at PL5 down at turn 8, then it's possible that this could be an answer; if a bit unintuitive for some.

    For it to work well total bonuses would need to be adjusted downwards (probably half tension bonus rather than full tension as the base of things): 1d8 becomes the die base. At or after turn 5, add a die. At PL4+ add a die.

    Just a thought.

  2. My concern is not so much what it does to the average roll but that it lets Bosses and PCs alike die to lucky stray shots in rare circumstances. The former might be exciting or disappointing, depending on who you ask, but the latter is almost always disappointing.

    But making it scale instead of providing a flat bonus is an interesting way to look at it. I don't know if I'll use it, but it is food for thought. Thanks for the comment.

  3. Do multiple instances of extreme terrain stack?
    If a mech is standing in a pool of acid, got hit by incinerator and is targeted by fire at will does it roll 3 times at the beginning of the turn?
    What about from the same source, i.e. 2x incinerator?

  4. What about electro sapper pods extra effect? Does it stack? If not which instance applies? First, second, better?

  5. Terrain Effects do not stack. This is covered in the Troubleshooting section.

    Sapper Pods do stack. To make this more obvious in the future their bonus Damage.will happen immediately.