Today, James Russel from Creative Assembly gave a conference about the origin of the Total War franchise and how Creative Assembly goes about developing a Total War game from the ground up.

He opened the conference with the statement that “Design is in the details”. And that’s exactly what the Total War franchise is about. Total War was initially created around the idea of creating a new kind of strategy game, which they successfully achieved by combining a turn-based campaign map and a strategic realtime simulation of the battles.

The Total War franchise relies on several key ingredients that all contributed to its success:

– Rich and varied world (give context and narrative background)
– Rival Factions (Either for AI or Multiplayer)
– Technology race (Gameplay Advancements through research)
– Non-military gameplay features (Economy, Diplomacy, Intrigue)

These are the key ingredients used to create a Total War game, but when you boil it down to gameplay mechanics Creative Assembly follow three filosophical pillars.

The first pillar stand for deep gameplay, James describes this by quotes Cid Meiers: “A game is a series of interesting choices”. This is what deep gameplay means to Creative Assembly. Interesting choices should always have pros and cons, for an optimal decision is a boring decision. This does not mean that a situationally optimal decision is bad.

The second pillar James discussed is the combination of a turn-based campaign map and real-time battles. The real time battles fuel the player with adrenaline as thousands of units clash on the battle field. Watching your enemies flee after winning a battle is a very rewarding experience. The discussion of real-time vs turn-based pops up, but James recognized that both sub-genres have its perks. It all comes down to the nature of the experience, which sub-genre fits the desired pacing of the game? Creative Assembly was successful as combining these elements to make a diverse and fun experience in which players experience various parts of warfare.

The third pillar is historical accuracy. All the historical setting give context to the game, providing the player with both interesting visual and a game world they can easily believe in. Creating a situation a player wants to invest time in is what strategy games are truly great at. But James didn’t have time to go into more detail on the historical part.

The next term James questions is reality. What is reality to a video game? He proposes a difference between reality and perceived reality. True reality is not something you want to aim for, because it would introduce many complications as Creative Assembly found out during Napoleon Total War, which I’ll discuss in a bit. But reality also has its pros, for people can use a lot of the intrinsic knowledge in games that follow realistic patterns. In the Total War franchise players can apply traditional war strategies like flanking, which will still deliver the expected result.

Another mechanic that is based on realism is the morale system, but it became something much more than just an addition to the realism of the game. The morale system adds a layer of depth and choice to the real-time battles; will you let the few remaining enemies escape or will you hunt them down, risking your own soldiers’ lives in the process? The morale serves other gameplay related reasons, such as to avoid mopping up phases at the end of a battle, allowing players to skip the boring part. Battles should never become a drag, so morale was a nice mechanic to cut a battle short once it gets boring, but Creative Assembly still allows players to hunt enemies with no morale down if the player wants to.

But the realism can also create problems, as it did in the Napoleon installment. When Creative Assembly introduced naval combat they ran into several problems with the physics system. Because in the Total War franchise, every single bullet is processed separately, creating a full feeling of simulation. The bullet physics also affect the cannonballs of the navy ships, so instead of flanking being most effective… firing at a ship head on did most damage because of the cannonball trajectory traveling through the ship for a longer period of time. Other weird things happened such as cannonballs from miles away did a lot more damage than a point blank assault. For gameplay this is not ideal, so they altered the reality to make it more enjoyable.

Another interesting topic James brought up was “Mystery vs Clarity”, revolving around the designer’s choice on how much information to show the player. But to show the player all information and variables in the game might not be the best idea. It’s true that players need to be aware of the basic statistics, especially new players. But do you want to show the more advanced variables? It comes down to basics vs powerplay.

James described the gameplay behavior of a powerplayer as the following: “If I can crunch the numbers, then I must crunch the numbers to win.” He goes on to say that dedicated powerplayers will go onto the internet to find out all the individual variables, and hes OK with that. The internet is a good place for that, if they really want to know everything. But by not showing all the numbers in the game you create a sense of mystery, which in turn leads to immersion and exploration. Let the player find answers, explore several possibilities. This all comes back to the first pillar, giving the player interesting choices.

James finished off the conference with a small list of tips designers should look for when designing strategy games.

The first tip is to consider game design as creating a bridge. Anyone could draw sketches for a bridge, even if it looks like a child’s drawing. But to truly design a bridge you need proper planning and schematic designs.

Start early with those schematic designs, if your programmers are still busy setting up the engine don’t be afraid to use Excel to create spreadsheets of all the variables and how they relate to one another. Creating the Excel sheets early on will reduce the initial amount of iteration once the engine is up and running. Of course you should still continue iteration, but you can do more focused iteration early on.

Create stock & flow charts for strategy games. Yes, charts are your friend. Stock & flow charts are clear graphs that depict the income and outcome of a variable or mechanic. An example from Total War is Gold. The income from gold is taxes and the outcome is army upkeep. So you get a basic graph that goes from Taxes -> Gold -> Army upkeep. But there are more influences, such as recruitment, which costs money but is a one-time investment. It’s good to list all these things that influence your variables and mechanics before you start working in a game engine.

Another thing you can work on in advance is the feedback loops. What sort of processes will the players go through, and when will they require these feedbacks? Try and find a balance between positive and negative feedback loops as well. If you consider these loops early on you can add surprising amounts of depth into the game.

Once you have an engine running you can start the real iterations inside the game. But beware of the dangers of tweaking… because game systems interact with each other. In strategy games, changing one variable can have many unforeseen consequences. So instead try and balance everything mathematically first and then look at ways of making them more fun. Think of how your system should interact; should a mechanic grow exponentially, fluctuate or diverge from the math.

And the last tip was the always remember the experience. This may sound obvious but if you’re working with complex game systems it could well be possible you lose track of the desired experience. Think of what the player should be able to do and how it should make him feel.

That was the massive summary of James Russel’s lecture on Total War’s design approach. I hope you found it useful, it was the most impressive lecture I experienced during GDC this year.