Last week, we introduced one of the new enemies in the expansion to Four Days: World Defense and talked about some of the mechanics of a tower defense game and the decisions we made in balancing ours. This week, we continue that discussion.
Consider Mortar Towers. Much like Flame Towers, they can damage multiple invaders at the same time. But unlike Flame Towers, Mortar Towers stack very well. Putting two Mortar Towers next to each other doesn’t decrease the effectiveness of either tower.
Mortar Towers have low damage efficiency, doing relatively moderate damage with a slower rate of fire, but their cost efficiency can be very, very high. In fact, they have the potential to do more damage with one shot than any other tower. A Mortar Tower does “area of effect” (AoE) damage, which means that it causes damage to every invader near its target. If the player can group a bunch of invaders together, a Mortar Tower becomes that much more effective. And that’s where tower synergy comes into play.
Disruption Towers are a standard AoE slowing tower. They put out a pulse that will slow down any nearby invaders. This has the obvious effect of keeping invaders on the map longer, slowing their exit and giving the player more time to kill them. It’s also great for setting up “kill zones” on the map where the player can upgrade lots of towers to maximize damage while the invaders are slowed down.
But Disruption Towers have another notable effect: they bunch invaders up. This was annoying for us during early tower balancing, when bunched invaders were actually a major problem. Consider a line of 10 invaders walking through a group of Gun Towers. If the invaders are far enough apart, the Gun Towers have time to kill each individual invader before the next one walks into their range. But if all 10 invaders are bunched together, the Gun Towers will be busy focusing on a single invader while the other 9 walk past unharmed.
The idea that Disruption Towers could actually cause this type of bunching was very unappealing, and it was difficult to justify their use in our early testing. Fortunately, bunched invaders are where Mortar Towers are at their most devastating. With some tweaking, Mortar Towers became a very good solution to the bunching issue, and the Disruption + Mortar combo became another interesting option for tower placement.
Essential tools for tweaking
There are a lot of numbers to adjust in a tower defense game. Changes to a tower’s range, damage, rate of fire, slowing effects, damage over time effects, or AoE radius can all have a major impact on the balance of the game. In our testing, we’ve found damage tracking/reporting to be absolutely essential. During development, we log the amount of damage done by each type of tower. We also log the amount of “overkill” (wasted damage beyond an invader’s HP) to gauge damage efficiency. These stats are printed out at the end of every game, and they’ve been indispensable.
But what about utility towers like the Disruption Tower? They don’t do damage, but their slowing effect can have a major impact on the game. How do you gauge their effectiveness? Are they slowing invaders too much? Our other favorite technique for tweaking game balance can help to answer these questions. Four Days has a replay system (only enabled during development) that logs every action the player takes. We save these actions to a file, which can then be used to automatically play the entire game over again with the same sequence of actions.
Even better, Four Days (like most tower defense games) has a “fast forward” mode for increasing the speed of gameplay. For players, our fast forward mode runs at 2x the regular speed, but for our own replays we run it at 40x. So tweaking towers is often an iterative process, even with Disruption Towers: we make a minor change, run the replay at high speed, look at the damage report, see how many invaders are making it through kill zones, and adjust accordingly. By doing several iterations, it’s possible to gauge just how much the slowing effects of a Disruption Tower change the overall balance of the game.
Interestingly, pretty much every tower has been “the best tower in the game” at some point in our testing. We want every tower to be great, and we want it to be satisfying to place each tower type on the grid. But there’s a fine line between “great” and “overpowered,” and we’ve crossed that line many times. Replays and damage reporting have been invaluable tools for recognizing when a tower is too good, and they’ve been just as helpful in determining how far to scale the tower back without making it feel underpowered. But we’ve also learned that, sometimes, a tower’s perceived weakness is actually one of its strengths.
Skywatch Towers and targeting
I never thought it would be beneficial for a tower to be air-only or ground-only. The more things a tower can do, the better it is, right? Imagine my surprise when it became clear that Skywatch Towers are good, in part, because they can ONLY hit flying invaders. There are 3 other towers that can hit things in the air, but the player is better off putting a few Skywatch Towers in the mix. Why? Let’s talk about targeting.
In my experience, one of the greatest sources of player frustration is a tower’s choice of targets. “No!”, they scream. “Shoot that guy! He’s almost dead! Why aren’t you switching targets?!” It’s difficult to find a targeting mechanic that will make every player happy, largely because the player’s criteria is “shoot the thing I am wishing you would shoot right now,” and we can’t easily write an algorithm to do that. We considered lots of targeting criteria and settled on “shoot the guy who is closest to an exit.”
Think of a long row of towers, where invaders have to march along one side of the towers, wrap around, and then walk along the other side of the towers to make it to the exit. The most frustrating thing these towers could do is focus on the newer invaders while letting further-along invaders walk right out the exit. By focusing on invaders who are closest to the exit, we’re choosing to minimize this particular type of frustration.
We also opted not to let towers switch targets. Once they lock on, they keep shooting until the invader dies or moves out of the tower’s range. This was mostly a performance optimization, since searching for targets is a fairly expensive operation. But we don’t treat it like an optimization — we treat it like a predictable game mechanic that the player should expect and plan for. Slow, strong invaders will inevitably draw fire from lots of towers, allowing weaker invaders to run past without issue. In many genres, this mechanic is called “tanking,” and it’s a big part of Four Days.
And that brings us back to Skywatch Towers. They’re the only tower that can’t target ground invaders, which means they’ll never be busy firing at a strong ground invader while a bunch of flying invaders zip past overhead. Reaping and Rocket Towers are useful for defending the skies as well, but you never know when they’ll be busy fighting something else. Skywatch Towers can still encounter tanking, of course, because there are slow, strong flying invaders too. But at the very least, Skywatch Towers provide a reliable source of air-only damage.
Reaping Towers revisited
So let’s return now to Reaping Towers and their need for a balance boost. It’s tempting to increase their damage, their range, or some other basic stat in an attempt to make them more useful. But those are drastic measures for a tower that already feels balanced in the early part of the game. It’s the tower’s late-game appeal that really needs to be improved.
Jumpers may feel like a bandaid fix at first — an invader that is best dealt with by using Reaping Towers, so Reaping Towers will be more necessary. But consider this: Jumpers actually give the Reaping Tower a greatly increased chance to get killing blows later in the game, and that’s a very intentional part of their design. Our hope is that Jumpers will keep Reaping Towers powering up throughout the game, fulfilling the Reaping Tower’s promise of efficiency and taking some luck out of the equation.
Thanks for reading. If you’ve enjoyed this article, be sure to follow us on Twitter and/or Facebook to be notified of future blog posts!