Deep Dive: How survival battles balance the demonic chaos of Total War: Warhammer III

Author: by James Martin, Creative Assembly

Game Developer Deep Dives are an ongoing series with a goal of shedding light on specific design, art, or technical features within a video game in order to show how seemingly simple, fundamental design decisions aren't really that simple at all.

Who: My name is James Martin, I'm a battle gameplay designer at Creative Assembly and I work on Total War: Warhammer III. My journey to where I work now started when I quit a previous job to say home and play Total War: Rome II on its release. A month or so later I joined QA at SEGA to work on Total War: Rome II DLC. I eventually moved to Creative Assembly to work on everything Total War.

What: Today I'm going to talk about the new survival battles we're working on for Warhammer III, and the learning curve they brought with them.

In previous Total War titles we've had multiple battle types; open-field battles, battles over settlements, full-scale sieges, and other kinds unique to certain settings. By and large we've steered fairly close to what we know. Our players are used to that familiar Total War formula which we try to maintain in each of our titles. Our games already focus a lot on asymmetry between various playable factions and the Warhammer series has the most unique take on this. Contrast armies of men fighting other armies of men in the classical world with hang-gliding goblins being launched by catapult into giant, ice-breathing elemental bears.

Total War: Warhammer III is a new title, it's the final chapter in the series, and we wanted to bring interesting new gameplay to our players while staying true to what constitutes a Total War title. Our aim was to create a really immersive battle experience for certain key moments in the campaign narrative which occur in the Realms of Chaos.

Survival battles are ambitious, with multiple new features that break out of that traditional formula. They're all about fighting your way across vast battlefields as the gods of chaos throw wave after wave of daemonic enemies at you, culminating in a boss battle against a greater daemon. You gather supplies en-route, which enable you to replenish, upgrade, and recruit new units to your army, and build tower defense-style missile towers and barricades to create bottlenecks, killzones, and deny passage.

It's a tough thing to tackle as a developer -- when you read back into the lore of Warhammer, the sheer scale of some of the battles portrayed is beyond what most of us could imagine, let alone represent in a game which needs to run across multiple PC specs. Bringing this level of scale to a game is very appealing, but it's also a scary concept to get right. You want the player to feel like a hero as they take on true hordes of daemons, but you must consider what level of scale is immersive enough, versus a scale which becomes unmanageable for the player with our current battle gameplay and feature-set.

We prototyped this mode early on in production to see how it played out and it was instantly a lot of fun. The mode riffed perfectly with the daemonic armies in Warhammer III and that was a big plus. The most interesting part was watching how people approached these battles in different ways. Most people thought their way was the obvious way to do things. Even with the same starting army, we found that people would take very different approaches, whether it was to focus on ranged units, artillery, tower-building, or spells. There's no single, correct way to play a survival battle, and we felt that level of choice was a good early outcome.

We planned to use a survival battle for our first public showing of Warhammer III, so we wanted as many hands playing it internally as we could, and not just developers -- people from all across the studio. That gave us lots of useful data. Peoples' success in the battle stretched from one extreme to the other, which introduced new balancing needs. Micromanagement has always been the challenge in Total War battles, and with the scale of survival battles, you'd think we'd be pushing players even further.

Why: The idea behind adding so many new features to an already brand-new mode was actually that we wanted each player to be able to find their own sweet spot in terms of micromanagement. The battle currency you generate can be used across all its features, and it's completely in the player's hands to choose where to spend it, and therefore decide what level of micromanagement they wish to employ during the battle. One player might want to focus on recruitment and upgrades for their army for example, and exert fine control over masses of warriors. Another might build near-impregnable hardpoints with barricades and towers and turtle up.

We also chose to split these survival battle maps into several stages, which 'unlock' as the player makes progress. This allows players to work in single areas at a time, and gives an important sense of focus in such chaotic environments. It also helps to telegraph where the final sequence of the battle is going to happen.

Having so many units in the field, both player-owned and enemy, can be intimidating to both experienced as well as new players at first blush. There was a tough decision to be made here: we wanted to hit the desired sense of scale, but at the same time, we knew that the sheer number of enemy units we wanted to throw at the player would be deadly. The battle was also in danger of becoming a slog, as the time needed to complete the battle would be unacceptably long. We needed to balance immersion with manageability. 

So we set to tweaking the enemy forces. We reduced the entity count of enemy infantry units by half, while doubling the number of infantry units the AI sends at the player. This essentially keeps the same amount of entities in play but gave us that sense of horde scale that we were after.

But it was still extremely tough. So the next stage for us was to reduce the health of all enemy units. This aids balancing at the larger scale, but breaks the consistency with what players might expect from that unit in normal (ie 'regular' battle) circumstances, and seems at odds with the fact that you're fighting daemons on their home turf. Importantly, we didn't nerf their damage output. Although the enemy unit health is lower, players still have deal with huge quantities of units that can still fight as well as they would in any other battle type.

Unit quality upgrades and health/ammo/stamina replenishments were added to give the player's forces more staying power. The abilities we have are simple to understand and are there for players to quickly make use of in the heat of battle. We also allow players to send units home to reinforce, and return them to the field when they're back to full strength. And if a unit dies completely, we allow the player to spend resources reviving them and returning them to the field, upgrades intact. We've never had a live recruitment and reinforcement system like this in Total War battles before, so it's a pretty exciting addition. It's there to allow for players to mitigate their mistakes (or bad luck), and enables a new style of play while supporting our scale ambitions.

Barricades and towers are conventional elements across many genres. We found that these really helped with lessening the load on the player. Barricades can slow or re-path advancing enemies and enhance frontline player units, while towers destroy enemy units on the approach to ease the pressure on the player's forces. It feels good to watch enemies getting whittled down as they come and knowing how it benefits you.

With such a high-stakes battle, we don't want to punish players who don't have the most intensive micro, but we still want to give hardcore players a rewarding challenge in a rich narrative setting. The asymmetrical nature of our factions, combined with the broad range of player approaches to battle in any setting, have made creating survival battles a challenge. 

If I were to start over, I would still be torn between keeping the context of these battles simpler in an effort to make the player-load and the balancing easier, and wanting to portray these crazy-scale conflicts against the gods of chaos. There's a tension between the two. But between the tools we give the player to manage the flow of the battle, and the considered series of checks and balances we've applied to enemy forces, we've found a happy middle ground that gives us our sense of scale in a deeply immersive, thematically appropriate and fun new battle type.