Dawn of Man is a compelling simulation that has players bring a tribe of humans up from hunter-gatherer culture through to the Iron Age. Part Civilization, part Warcraft and part Dwarf Fortress, it melds each playstyle together into something unique.

The game's developer, Madruga Works, was founded by game industry veterans Tucho Fernández and Martiño Figueroa -- Fernández was a 3D artist at Ubisoft and a contractor on many video games, while Figueroa worked on AI and gameplay programming at Criterion Games for franchises including Burnout and Need for Speed.

Here, Figueroa explains key decisions in developing Dawn of Man, and how the release of their previous game, Planetbase, informed their latest game.

Answers have been edited for clarity.

Making the leap to independence

[T]owards the end of 2014, we both decided we had enough savings to survive for a year without a job, and we decided to try to make a game and see if we could make a living out of this. I've always been a fan of real-time management games and we decided to do something in that genre, but a bit different: something with more involved AI, a game where you would build some sort of "ant colony" and could then watch your people live in it.

We came up with the most minimal design we could, in order to finalize the game in less than a year, and this was how our first title Planetbase was started. Tucho did all the 3D Art, I did all design and coding, and then we got a few contractors (which we paid with our savings) to help out with audio, animation and UI art. My old colleague from Criterion, Tom Williamson, did the Xbox One/PS4 port. Planetbase was a massive success for us: allowed us to get back what we invested in it, pay ourselves for the work we did and fund our next project.

We learned a lot in the process of making Planetbase, and wanted to have another go at improving our particular survival/city builder recipe, but we wanted to try another setting. We were always fascinated by that time in human history: a bunch of people with sticks and stones in the middle of a forest somehow managed to make a living, survive and evolve through the millennia.

Both games are similar in that for the most part it's all about producing/gathering resources to keep your people going, designing an efficient base/settlement and juggling the various production chains. However while in Planetbase the management was mostly at the macro level, in Dawn of Man you can give commands to your people with varying degrees of precision. As an example: You can tell them to a) "gather sticks in an area until a limit is reached," b) "gather these particular sticks," c) "tell this particular guy in the settlement to gather that particular stick."

Our favorite city builder/management games are probably The Settlers 2, Stronghold, Startopia, and more recently Banished, of course, and Frostpunk is great too. Dawn of Man also draws from RTS games like Age of Empires or Starcraft.

However, we believe our games have quite a unique mechanic: autonomous AI that constantly has to make life or death decisions. We really haven't seen that in any other game. It's a lot of fun in our view to design your settlement, give the people a few basic directions and then see them trying to figure it out. This also has it's problems of course, as sooner or later someone will make the wrong call and get themselves killed, but we believe this is part of the fun too.

Everyone might have their preferences here, but we always believed that games that track the actions of individuals are a lot more fun, otherwise you end up with some sort of spreadsheet-based gameplay where everything is decided by some math behind the scenes.

If you send three people to get wood you should be able to see them go to the tree, cut it down, and bring the resources home. And if they have a chance encounter with a cave lion on the way back then it makes it even more exciting. This of course has the disadvantage that it is a lot more effort to create gameplay, because you need to create each individual interaction, which is why probably some games choose to use a simpler approach to game mechanics that allows them to generate a lot more content faster.

Dev tools and process run-down

From the programming point of view, Dawn of Man was created in Unity and coded in C#. We created our own shaders, to realize our own particular visuals and atmosphere and customized the Unity terrain system quite a lot in order to be able to generate the season transitions, add/remove trees and place removable resources.

Artists used a variety of tools to create 2D and 3D content: Maya, Blender, Topogun, SpeedTree and Zbrush for 3D assets. Photoshop, XNormal, CrazyBump, Substance Painter and others to create 2D textures.

And yes, we are working on the PS4/Xbox One version of the game, we are redoing the UI and control scheme to ensure the game is easy to play with a pad, which is always a challenge for these kinds of games. We have some experience doing this with Planetbase so we hope to be able to offer a quality port.

We have a very particular way of working at Madruga Works: Tucho and I build a prototype of the game first, then once we are ready to go into full production we get remote contractors to help us finish the job. They are all usually very experienced people that have worked in the industry before, many of them old friends or colleagues. In Dawn of Man, there were a total of 10 people involved, about half of them 3D artists.

Regarding the generation of assets usually we start by searching for references to have a general idea of how the asset would look. We then create a placeholder and put it in game and adjust the proportions. Once we are happy with the proportions, we make a quick concept art (usually overpainting a capture of the placeholder) and then we make a high-poly model using this as reference. We use mainly Zbrush for this, but some of our contractors also work with 3D-Coat at this stage.

Once the high poly model is done, we do the re-topology to get something usable in-game, we use Topogun or 3D-Coat for this and then, we bake the maps from the high poly mesh and paint the textures in Substance Painter, but as we switched to Physical Based Shading during the development, we also used a bit of the old school tools like XNormal or Photoshop for this.

History fit for simulation and management

From the design point of view, my goal was to get the player to experience the great transformation in the way humans lived in the various eras, and get them to adapt their play style to that.

First, you start off as a hunter gatherer band, which makes it really hard to sustain a large population. You have to constantly send hunting parties to get food, which decays quickly and can never be stockpiled for long. Then when cereals were domesticated society was transformed forever, humans could then grow their own food, and store it for long periods of time (grain if stored dry can last for years). This allowed them to settle in much larger communities, where a large part of the workforce is employed in farming.

Finally the smelting of metals was another huge change, as it allowed us to create much more effective and durable tools. As populations grew they started competing for resources more actively and human conflict was much more common, [so] you will need to actively fortify and defend your settlement.

One thing we've realized is that games are played by a very varied group of people. Some are looking for a chilled experience to play for a couple of hours after work, some are looking for a very challenging experience that tests their skills. Hardcore mode offers a more challenging experience by offering more realistic animal aggression, more raider attackers and only one savegame slot.

But they key difference is that it is a race against the enemy raiders, as they progress their technology at their own rate. If you don't hurry up developing your people, you can end up with your hunter-gatherers armed with sticks and stones being raided by 40 people carrying steel swords and composite bows.