Update #28: Update #28.
So it's now June and some of you were expecting a prelude release. We've been communicating our progress on our forums but we haven't done a Kickstarter update in some time. We're not quite ready for a prelude release yet but we have been hard at work!
We know many of you have high expectations and you've encouraged us to take the time to make a great game rather than rush to release something mediocre. We are putting a lot of care into those things that will make Sui Generis a truly unique game. Of all the games we've played it's those that pay attention to detail and have layers of subtlety that have really engaged us. While we can't satisfy your every request we do understand the importance of making something good, not just "good enough".
Until recently we focused mostly on outdoor environments which in many ways are the most technically challenging. Because we're such a small team with one programmer we generally have to focus on one thing at a time and the underworld had taken a back seat. Developing the prelude we found ourselves confronted with some issues that had been neglected and reviewing things that we weren't satisfied with.
You might remember we promised a vast and sprawling underworld, it's all interconnected rather than being many distinct environments. It's also very deep and it changes as you go deeper. The concept remains the same but our early design now seems too basic and not to the standards that the game is shaping up to.
Our first goal was to introduce more variation in the common structural graphics. We improved the modular system through which the environment is constructed and developed a tool that allows us to create these environments extremely quickly and interactively while providing an intuitive workflow, this means that we can focus on and experiment with design rather than getting bogged down in the work itself. We've improved and adapted our existing assets to this new system and we're developing new ones with fresh insights.
Unlike the terrain the underworld is not a single streaming environment and there are some very different travel mechanics. The underworld regions are quite large and complex, the first "level" of the prelude contains almost 10000 meshes, 100 shadow casting dynamic lights and many characters so we need some important engine features to be in a very much finalised state to drive all this. A lot of this work is done now but there are still a few important issues to resolve, especially for lower end systems.
One really important issue we struggled with for this type of environment and the isometric style view is visibility. Being able to see past walls is very immersion breaking, it removes depth from the environment and also causes difficult problems with balancing line of sight for the player and AI. This is especially true when AI has complex behaviours and may be interacting with the environment. The approaches we've seen used in other games would either not work or be visually disturbing.
Finally we had a promising idea, prototyped it and instantly loved it. What we're doing is dividing up the environment based on doorways and only allowing you to see the room you're currently in or any room that you're on the threshold of. The rooms fade in in such a way that as you move around you barely even notice it happening, the visual impact of the environment and exploration are much more immersive.
AIs can be subjected to the same exact visibility rules as players putting them on a level playing field. As a byproduct we've achieved a sort of old-school encounter mechanic which we really like. As you approach a room you are able to see what's inside and what's inside is able to see you.
This just naturally ties in with other mechanics and how AI understands and navigates environments. For example if you were being pursued we can accurately determine where AIs see you go and where they might search if they lost sight of you.
One of the things we care about most is persistence, we want everything in the game to feel substantial and real. Everything can be interacted with and the results of every interaction are lasting. To support this the bodies of the dead must not just dissapear but remain in place until someone moves them. These bodies must therefore slowly decay. Necromancers must also find real bodies to raise, they can't just will them out of thin air.
We're leveraging the power of our recently improved procedural character system to make bodies gradually decompose, that of our procedural animation to make their movements uncoordinated by various degrees. Combine this with some great voice effects and unusual AI and you've got some pretty compelling undead.
DAMAGE AND EQUIPMENT MECHANICS
All the mechanics for armour and weapons are now in place and working as we wanted and better. There are four basic damage types: slash, pierce, crush and impact. All of these behave differently in terms of how the forces of collision translate to damage, for example slashing damage is effective when cutting and even with weaker forces whereas crushing damage requires significant impacts. Base weapon stats include these damage types and also various properties which control how they behave physically. Weapons are not just more or less powerful but each weapon feels and behaves differently in combat having a huge effect on how you play.
Armour has mitigations for the four basic damage types and also other properties. We are going for very realistic behaviour for armour, plate armour is effectively proof against bladed weapons and extremely resilient to all damage types, chain offers great protection against some damage types but not others, common leather apparel is pretty useless as armour. There are various other types of armour each with distinct qualities and things get interesting when you realise they complement eachother. It is entirely possible to wear chain under plate, or a gambeson to absorb some impact and effectively reduce the force of blows to a point where the plate can mitigate them completely. When you equip some armour you really get a sense of being armoured and protected, you will even hear the plates sliding over each other, the chain jingling and the leather creaking as you move.
We've extended our inventory system to support 8 layers and over a thousand slots, we don't use them all but they give us the flexibility to model the complex interactions of our growing repertoire of wearable items. All armour is locational, it covers different parts of the body and with different efficiencies, layers of armour interact locally on what they cover. Each part of the body also has varying base vulnerabilities to damage types with different effects. Head blows can disorient you, leg blows can throw you off balance, vital organs, bleeding etc. are all taken into account.
The best thing is that you can see, hear and basically feel all of this happening as you play the game. Weapons have realistically distributed mass, swings make swooshing sound based on the weapon and its current velocities, there are dynamic collision and other sounds, procedural animation behaviours, different cries of pain and sprays of blood which all accurately reflect what is happening. All these things are possible because there are measurable physical forces and interactions behind everything. They're not just effects on top of some simple mechanical system but a result of the simulation.
The combat and animations have been evolving constantly, they are the thing the game has been most praised and criticised for. Making notable improvements has been extremely challenging, it's a hugely experimental and difficult technology and we're often shooting in the dark, sometimes missing and wasting time.
Originally we designed the combat with a big focus on the physics, the force of blows and movements of your character being very dependent on the fluidity and timing of inputs. We'd had a lot of practice with this but when we released the combat alpha we saw it wasn't always well received. Many people perceived it as somewhat unresponsive and clumsy, watching videos of others playing the game we saw that even skilled players weren't easily achieving the same harmony with the physics, their character's motions often looked ungainly. Elegance and effectiveness didn't necessarily go hand in hand. We quickly made changes to improve the responsiveness of the controls based on feedback but these hurried changes came at the price of some unrealistic behaviour.
So, having received a lot of feedback on the combat and prototyping various suggestions in several rapid releases, we then went back and revised some of the earliest and most fundamental aspects of the animation system. We shifted the balance away from pure physics towards player control, completely redeveloped the input system and fixed numerous issues that were present in the combat alpha. We also made improvements to the AI which was still incomplete, giving the AI more tactical and human similar behaviour and trying to counter potential exploits. Our redesigned AI proved extremely difficult to beat which is what we were aiming for when considering the best possible AI opponents.
This took some time but as it all came together we finally found ourselves before what feels like a new game. The many changes to animation systems and a new asynchronous processing of inputs led to very responsive character control, characters move fluidly and remain balanced even under the most extreme conditions, allowing for precise control over new attack combinations and very fast dodges. Our almost unbeatable AI became easy to defeat and we had to add a whole new layer of cleverness and precision to it, finally making it challenging again and then introducing skill levels to reduce and control it's effectiveness.
The final result is fantastic, it looks great and it's incredibly good fun. The foundation is now more solid allowing us to do more, we see ways to improve it further but already we've achieved better results than we initially thought possible. Looking back and seeing how far we've come since the early animation we wonder at what the future could hold!
These are just some of the most important things we've been working on, we have been consistently achieving our goals and the results keep exceeding our own expectations. The prelude is still in the works but we're very close to a first alpha release, we need to solve a few issues but further releases should be close behind. We're extremely pleased with the core combat in general now so after a quick final alpha test we plan on also releasing a beta combat demo. This kind of gameplay will also feature in the prelude in the form of an arena mode.