Howdy Folks!
Hope you're all having a pleasant beginning to your August. As my vacation is coming to a close, I figured it would be a good time to prepare you all for what's in store for H3 for the remainder of the Summer.
We're going to be starting a new Alpha cycle, similar to what was done for Update 52 (but hopefully much shorter). For those of your not familiar with the last cycle, in short, Updates over the coming weeks will ONLY be pushed to the Alpha Branch on steam.
In addition, during the Alpha cycle, several things will be different about the game that will result in a sub-par experience for folks used to 'things as normal' such as:
- Some scenes will be intermittently disabled or broken without warning as core systems are replaced/modified.
- Game balance will be WAY OFF in modes like Take and Hold, while things related to firearm damage, etc. are rebuilt.
- Performance may be terrible in certain scenes, with certain objects, agents, etc.
- Bugs, so many bugs
Now, you might be asking, why is this all happening? Well similar to with Update 52, there are systems in need of work, to support future modes, to improve performance, to fix long-broken issues, that need a level of rewrite that is... not frictionless. Some of this work is experimental/iterative where testing will need to be done in phases to determine if design decisions being implemented are the right ones. It will likely get messy. For the Update 59 Alpha Cycle, my focus points and goals are as follows:
System Set 1: Ballistics System Refactor
H3's current ballistics system is OLD. Much of its structure and thinking actually predate the game, and come from an earlier non-vr prototype I was working on, back when... I... um... was a far far worse coder. It's wasteful, tries to simulate far more than is needed, is almost impossible to control design-wise, makes bad assumptions, and doesn't actually produce terribly interesting ricochet and penetration behavior, despite performing a lot of expensive calculation. It was also designed at a time where the player was the only one firing projectiles, and so its computation overhead was acceptable. That is no longer the case so...
SO, instead of continuing to do what I've done for the past 2.5 years, which is just to throw more on the pile, and engineer around old terrible decisions, I'm ripping off the band-aid and completely rebuilding the projectile class itself, and the generic damage class system that everything uses for transmitting 'damage' and related things like force, references to things that do damage, etc.
The hope is that when this is all done, we should get the the following benefits:
- MUCH cheaper projectile code in general.
- More interesting, differentiatable behavior when projectiles of various types strike various materials. Wood, plastic, and thin metal, plaster and like will be more able to be penetrated.
- Ricochets will be more common and occur at the types of angles they should.
- Multiple penetration will actually happen when it should.
- Transmission of damage, especially differentiating armor-penentrating rounds from expanding and frangible projectiles will be more interesting and accurate.
- Projectiles and melee hits will be able to 'carry' other types of damage as well, from thermal, to emp, to just physics force. Instead of a nasty hacky set of additions i've made over the past year, this will all be unified.
In short, it will matter more what type of round you're firing out of a gun (AP, FMJ, JHP, Slugs, Buckshot, Flechettes, etc.). Certain round types will be far more situational, and there will be more reasons for me to to make combat environments contain more types of materials, as the penetrability of those materials will make the range of cover-to-concealment more tactically interesting based on weapon platform.
It will be a cleaner, faster, and roundly better simulation, and despite juuuuuuuust how many things I will have to break and rebuild to make this happen, it will be more than worth it in the end. The goal is to make this the last time I ever have to significantly change or expand this part of H3's systems.
System Set 2: Impact Sounds
This was originally planned for Update 52, but due to the changes that were needed both for core Physical Objects and Melee weapons, I opted to hold off on this upgrade rather than delay Update 52 another 2-3 weeks. So we're going to finally deal with this now. The focus will primarily be on cartridges, shells, guns, mags and melee weapons, but other generic holdable objects should also get a sonic upgrade from this. Everything will be migrated over to the sound pooling system, so in addition to sound better, impact sounds should get the consistency and performance improvements brought by those systems.
System Set 3: Sosig Agents
Anyone who has been watching the Twitch streams know about these. I've been planning these on paper since the New Year, and development on them is finally moving apace. Essentially a Wiener-bot Mk. 3, the Sosigs are the first truly new bot approach I've taken since the first multi-segment agents.
They will have a host of new features including:
- Truly physics reactive movement and behavior rather than the current bots just pushing through things
- Ability to 'crouch' and go 'prone' to more effectively use cover
- More efficient and reactive pathing that will allow for the placement of obstacles without the bots freaking out
- Complex damage system that includes leaking mustard, joints that can be broken, the ability to fall down and recover.
- Jointed weapon holding such that held weapons can be deflected and disarmed
- The ability to be grabbed, whacked, thrown, stabbed, lodged, and every other manner that you folks enjoy beating your meat.
- Significantly more nuanced and complex perceptual system that will be able to be manipulated better with toys, sounds, as well as enabling stealthier game-play
- More hilarious Voice Lines
- HATS!
The order of development on these will proceed from the ground up in three rough phases:
Phase 1 (in progress): Sosigs will be implemented as incredibly reactive targets first, with damage balancing done against firearms, melee weapons, explosions and physical/collision trauma.
Phase 2: Following this, they will be 'taught' to 'fight locally', meaning to react to targets in their immediate area and attempt to dodge fire when engaged. Their only navigation in this mode will be immediate vicinity, and they won't yet 'know' anything about the map in general
Phase 3: A new, much more efficient pathing-and-cover grid will be built for them (available only in some modes), that will allow them to cross large distances to fight, pick cover and engagement direction based on what they're fighting, flank, and run away. This will be the _hard_ part. Truth be told, I'm not even sure this will work out how I envision, but I'm going to give it my best shot.
Important Note on these: Due to just how different these bots will be from the existing wiener-bots, these will NOT be replacing the bots in Wurstwurld and Take & Hold. They will behave far too differently such that those entire scenes would have to be rebuild/redesigned with them in mind, and that effort is better spent on new content and modes moving forward.
The Update 59 Alpha Cycle Will Begin On Friday
If you would like to participate in the Alpha: Right click on H3 in steam, go to properties, go to the betas tab, and switch to the branch called 'Alpha 1'. REMEMBER, things will be broken randomly during the Alpha cycle, and there may be multiple 2-3gb downloads a week during it. If bandwidth usage is important to you, do proooobably don't want to be on the alpha.
If you do NOT want to run the Alpha: Do the above, but make sure you have 'none' selected in the betas drop down.
So in conclusion, to those of you who stuck through reading all of this, thanks! and I hope you're excited to journey through another Alpha cycle with me. This one's tremendously exciting to me, as its emphasis is on truly refining the core game-play systems of H3, creating new opportunities both tactical and ridiculous, and establishing the ground work for some exciting seeeeecret things coming early in 2019 (taps fingers together maniacally).
Hope you all have a wonderful week, and I will see you all on stream this week, or on the devlog on Friday!
Peace,
Anton