The Outsiders started work on what would become known as Project Wight in September of 2015. For the year and half preceding that date, David Goldfarb (Chief Creative Officer) and I had been working on pitches to publishers and investors to get the cash required to start development.
What was different about this project compared to the 20+ others I’ve worked on in my career was that this time, before we had a team or any kind of business structure around us, we knew exactly what game we were making. No game jams, no brainstorming, no market research.
In that sense the demo you saw on stage at Unite or on YouTube, Reddit, Twitter or Facebook was a very accurate realisation of what David had envisaged for the game for many years beforehand. It actually almost perfectly matched the first concept art we had commissioned for pitches in 2014.
When reviewing this blog post, it’s important to understand the difference between our prototype (a 2 ½ hour long fully playable build focussed on gameplay that is not public) and the demo (a highly polished visuals-focussed ‘golden path’ through that prototype that we showed at Unite and in video).
So how did we get from this private vision in one person’s head to what many people feel was a successfully realised demo? There are three main high level directions we took to get there so efficiently.
Prototype to test game mechanics
As mentioned before, we knew what game we were making - we didn’t need to spend time figuring out a high concept or a killer game mechanic. What we did need to do was confirm which aspects of this design needed focus and which of Dave’s assumptions from a mechanical and gameplay perspective would work and which wouldn’t.
Because we needed flexibility and speed during this process we decided it was much better to prototype than approach this as a pre-production phase. When prototyping, code and artwork are created as fast as possible in order to test an assumption, with no concern about whether the code or artwork will stand up to close scrutiny or be shippable to the public. In a pre-production phase things are quite different -- developers are figuring out the exact method to build something so it is suitable for the final product.
We decided to allocate a year of prototyping time to the project, with a small team of 7 people (one Animator, two Coders, an Environment Artist, a Character Artist, David and I)
Footage of the Project Wight prototype in September of 2015, after one week’s work
Stick to the vision
In the course of prototyping a game it’s very tempting to keep things very open, to the point that key mechanics, player perspective or even the genre of the game itself can change during the process. Because we had such a positive response to the high concept of the game during our pitching process and when explaining the game idea to friends, we worked very hard to constrain the prototyping process to ensure that we retained this already proven high concept.
What this meant was we never considered some things that other teams might have explored in the same circumstances. We never tested a third person player camera, controlling a human rather than a creature, and we never looked at alternative settings other than caves and natural environments in a nordic, dark-ages setting.
What we did do however, was focus on answering questions related to key elements that we knew had to work in order to sell this concept. Our prototyping focusses included (but were not limited to):
What do we need to do to ‘sell to’ and ‘remind’ the player that while the game is first-person, you are controlling a creature rather than a typical human player character?
How big are these creatures (at various life stages) relative to their human enemies?
What kind of locomotion methods work to sell the creature feeling and to make navigating the environment compelling?
What size are the smallest and largest environments the player will encounter?
What is important to focus on when building a first-person melee combat system for a non-human player character?
Can we build compelling medium-term gameplay without explicit missions or a plot?
Footage of Project Wight in October 2015, after a month of work
Rapidly build an UGLY prototype using Asset Store content and tools
As well as having the discipline to ‘stick to the vision’ it was important to us during this prototyping phase to also work as fast as possible and not worry too much about building things correctly. We decided very early on to do ZERO work on the visuals of this prototype until as late as possible. We wanted as much flexibility as possible in world building in particular -- once you do an art treatment on a complex environment it becomes very difficult to do small or large tweaks to make the game more fun.
As well as this broad-brush ‘greybox’ approach to environment prototyping we had a similar approach to characters and other interactive elements in the world. In this case where the geometrical flexibility isn’t so essential, we opted to use the fastest method we could think of, which was to grab free or inexpensive asset store assets for those game elements. A great example - just six weeks before we showed Project Wight onstage at Unite our only human enemy model was still the human viking character from Unity’s Blacksmith demo that we had integrated in December 2015. If it ain’t broke....
Footage of the tree section of the Project Wight demo, two months before the game was revealed
Rapidly ‘Skin’ the prototype with high quality AAA assets
By early August 2016 we had locked down the features in the prototype and we ready to start turning what we had -- a very fun two and a half hour prototype -- into something that looked amazing and punchy enough for a two-minute on-stage sneak peek.
We knew we couldn't art treat the entire environment to demo quality, including all the gameplay and objects in it, so the first task was to identify what the ‘golden path’ was through this large amount of content that we wanted to show on stage and in video.
To help with this decision, we decided it was essential in the demo to communicate three key aspects of the game’s design:
That the player character in the game is a non-human creature with non-human abilities
That the game takes place in an alternate dark ages where humans and these creatures are in conflict
That the creatures in this game world become more powerful as they age
Identifying these three focus points made it easy to decide which elements of the prototype would be ‘skinned’ to AAA final-quality. We would start with the setup of the child creature seeing the slaughter of their parents and running away from humans; we would use environmental storytelling, AI and game mechanics to show the conflict in the world; we would have a clear ‘victim, then revenge’ turnaround in the demo to show clearly the variety in power and subsequent gameplay differences between different creature life stages.
Once we had identified the path through both the abilities and the gameplay we set about working as fast as possible on improving the visuals. The main areas we focussed on in this final ‘skinning’ stretch were:
Iterating the animations to move from functionality-focussed prototype movement to something highly polished
Replacing placeholder first-person creature models with final-quality versions including geometry, textures and shader tweaks
Replacing our free placeholder Blacksmith Demo human characters with some made by contractor Sven Juhlin (who had just completed work on Guerilla's Horizon: Zero Dawn)
Replacing Asset store and placeholder particle effects with some made by contractor Andreas Glad (who had just completed work on DICE’s Battlefield 1)
Replacing placeholder props and objects with final quality ones
Replacing placeholder sound FX and adding new audio where they were missing with the help of Redpipe Sound Design in Stockholm
Replacing the grey untextured blocked out cave walls and outdoor environment with our own photo-scanned objects (textured with Substance Designer) and with meshes and textures from Quixel’s Megascans library, the RTP Asset Store tessellation solution and landscapes built with World Machine
Carry out a final lighting, fog and camera effects pass using standard Unity lighting along with free Asset Store tools like the Atmospheric Scattering solution from The Blacksmith demo
Optimising the scene itself as well as doing code optimisations to increase the framerate at 1080p from 30 to 50fps
Footage of the project Wight Demo with two weeks to go before demo completion
So where are the team now? Well we still have lots to prototype, so we archived all the amazing visuals and are back working in a grey box environment for speed and flexibility. Sometime next year we start pre production. Unity has been a great tool for prototyping but it also lends itself very well to creating large, ambitious games with high production values.
Visit madewith.unity.com/theoutsiders for the latest information on The Outsiders and Project Wight.