Blog

Hi folks !

Let’s start by a quick update on our kickstarter journey. We received confirmation from Amazon that our account is valid and that our Kickstarter account is now linked. We also finished filming all the scenes for the trailer and entered post-production tonight. We still have to get Kickstarter’s official approval for our project, but it is coming together nicely.

Anyway, this week, apart from dealing with Amazon and work on the trailer, we did a lot of long term work on our systems. Mostly obscure and arcane thing like speeding up our data structures, cleaning up some code and revisiting our pathfinding algorithms.

This is François at the keyboard by the way. Germain is super-busy working on editing the trailer. He is doing an amazing job so let’s give him some space so he can keep doing it.

This week’s article is about one of my favorite parts of game design: combat systems.

Designing a good combat system for a strategy game is absolutely crucial because every aspect of the strategy will be influenced by how the conflicts are solved. If your system is so random your players can’t tell if their decisions are good or bad, you failed. If there is a significant chance that a single inexperienced footman critical hits his way through an army of mounted spearmen and into another player’s castle, you can be sure to lose your right to call yourself a strategy game. On the other hand, if your game is too predictable, you may end up having the same problem as the guy who invented tic-tac-toe: any mammal intelligent enough to try seriously can find the single unbeatable strategy in about 10 minutes. Fortunately, randomness isn’t the only way of achieving unpredictability. A good system (like Go, for example) can be completely unpredictable without being random nor having a complex system of rules.

Anyways, how can we balance and design a combat system? What sort of compromises do we have to make? In our opinion, there are three big aspects to decide while building a combat system.

1. What sort of balance do we want ?

Almost everyone knows about the rock-paper-scissor pattern in RTS games. It is the system in which each unit is strong against one unit and weak against another. The rock-paper-scissor balance is quite simple to understand and generally creates a gameplay about trying to predict what unit your opponent is going to produce. If you can find out or deduce what weapon he chooses, you can use the counter strategy and beat him. That sort of system is nice. It is a simple way to let people try to outwit each other while not having to worry too much about production costs and unit strengths. Everyone is happy and equal, and beginners can find themselves winning from time to time.

The numeric balance, as we like to call it, is a type of combat system where you could say that any single unit can be beaten by the right amount of another one. In other words, it is the “One paladin is worth two footmen” balance. It brings an interesting gameplay for players who like to make calculations and it also allows people to ask themselves “Can I manage my paladin in a way that will make him kill three footmen before he dies?” That sort system can allow for much more exploration and replay value. On the other hand, it demands a lot of testing and tweaking. And since it means not all units are happy and equal, the balance of your combat system will transfer some of its complexity to your production pipeline. For example, if your Kung-fu master is worth ten times your regular soldiers, he should be somewhere around ten times more costly to produce. If you need to have at three monasteries in play before you can start producing a Kung-fu master, maybe he can be just as powerful but only take eight time longer to produce, so as to compensate the fact that you need to create special buildings before you can produce your Kung-fu army.

The third and last pattern we know about is what we call the “Option and Combo” balance. It is what you generally seek to achieve when you create a huge amount of different units or give every unit a huge amount of options. It is often found on role-playing games where you have so many options you can never tell clearly what is best for you without trying them. That kind of system is a strange gamble because it brings a lot of possibilities for exploration and has a nice replay value, but it is also dangerous because it is hard to predict what kind of combos players will come up with. You may test your game during hours, you can never know what someone thinking outside the box will find out. Let’s say your Viking warrior has a fear ability that can be combined with some weird moving japanese temple with a surrounding slowing effect. Together they would make an incredible combo where you can trap your opponents into staying under your archers’ fire. You could not have guessed your players could do it, but now your game balance is broken.

Of course, I am describing that as if you had to choose only one of these patterns. In reality, it is more a list of basic principles you have to understand when building your system. You can have an overall numeric balance with basic rock-paper-scissor mechanics between the different categories of units. For instance, pikemen could be generally good against mounted units and weak against archers, but they could still be wiped out by elephant riders because they are more powerful and expensive than regular horse riders. Maybe you can have a simple rock-paper-scissor balance but spice it up by adding some nice special powers to players depending on their factions. That way, they can choose the right moment to use combos and turn the tide of battle.

2. How do you achieve that balance?

Ok, so now you know what kind of balance you want. Unfortunately, it is not the only thing to consider. There are many ways to obtain a certain type of balance in your combat system. I personally like to think of it as a compromise between numerical and contextual balance.

Numerical balance is plain and simple. If you want your pikemen to be good against horsemen, there is one simple way to do it: give the pikemen a statistical advantage on his dice-roll or a damage bonus against horseback units. You want the paladin to be as powerful as two footmen? Give him twice the amount of hitpoints or make him deal twice as much damage. Balancing your system by balancing numbers is easy and rock solid. However, it may sometimes be too solid. Strategy games are not always only about numbers and calculations. They may also be about tactics, opportunity, and reflexes.

Contextual balance is a bit more subtle. If you want your pikemen to be good against horsemen, maybe you can give them a bonus against charging opponents and make charging the main way horsemen attack. This way, pikemen are good as long as your opponent uses his charge attack, but if your are the one using horses and you see your opponent has pikemen, you can find a way around the situation and use your horses in a defensive manner. To make your paladin twice as strong as the footmen, you can give him a regeneration power that allows him to fight long enough to kill two footmen, but leaves him equal to them when the time comes to attack a structure. Another, perhaps even better example, would be the range advantage. What makes archers good against footmen is not a matter of numerical bonuses, it is the fact that they rain arrows on them and footmen are defeated before having the chance to fight. However, in some situations, when the footmen ambush their opponents, the range advantage of the archers can be nullified and footmen can actually be exactly what you need. Contextual balance allows for creativity and strategy. However, if everything is relative, your game can become too hard to predict and you may end up not being able to balance it at all.

3. How do you roll your dice?

Finally, suppose you know what you want and how you will achieve it. There is one last thing you need to understand. It looks trivial, but it is not. If there is a part of randomness in your system, you need to understand and be familiar with the concept of statistical variance. The way you create your randomness and the number of times you pick random numbers have a big influence on your system.

I know that modern computers and simple mathematics can allow for very fancy random distributions, with interesting noise patterns and lots of subtleties, but for now, let’s just pretend we are using dice like in a good old pen-and-paper roleplaying game. I will speak in terms of dice because it is easier to understand for a normal, non-mathematician human brain.

There are two popular dice paradigms in the pen-and-paper roleplaying world. The d20 and the dice pool. The first system produces a simpler random result. The players roll a single die and add something to it; the better you are, the bigger number you add. Such system is likely to produce extreme results because each action someone undertakes has equal chances of being anywhere on the luck spectrum. A bonus on the die rolled (or a bigger die) will change the long term results of the combat, but each individual action has a good amount of chances of having an extreme result.

The second system, the dice pool, is a system where people roll a bunch of dice every time they do something and the result is given by the sum of the dice. While that system doesn’t look so different, it is, because rolling many small dice produces a result that will statistically be closer to the average than rolling one big die. For example, if you need 10 on a ten-sided die, you have one chance out of ten. If you need 10 in a system where you flip ten coins and each tail gives you +1, you need to have all tails to succeed. By the way, that means one chance out of 1024. Chances are that you will get something around 5 tails. A dice pool system produces more stable results, that is, the more dice you roll, the closer to the average you can expect to end up. That creates a nice sense of consistency in the gameplay and can make lucky hits more special. If you are not careful, however, that can lead to a situation where a unit strong enough may actually win against hundreds if not thousands of weaker units. You may find yourself in need of some sort of failsafe where there is always a minimum damage on an attack or small constant chance of missing no matter how good you are.

How does that apply to a computer game now? Well, if you use randomness in your combat system, you may not use actual dice but it may be important to understand the difference between a linear system, which produces a uniform distribution and a cumulative system that produces a non-uniform distribution (or any other more complex system that would do something similar). Taking that into consideration may help you avoid a lot of frustration for your players or deliberately allow funny results.

In conclusion

After that long text about understanding and balancing combat rules, it all depends on your game. There is no perfect system and while a huge, complex and perfectly balanced system can be interesting, sometimes you just need something basic to let your player concentrate on the other complex and interesting things in your game.

In designing Castle Story, we tend to gravitate towards a Combo and Option system with context-sensitive rock-paper-scissor dynamics. Your units’ effectiveness will depend on the structures you create and the way you adapt the world around you. The actual fighting will be dice-pool determined, with a basic physics flavor.

Team Sauropod