Icon: Voidfall Demo · published

Alpha v0.1.11b is a smaller follow-up to our massive 0.1.11 release, focused on combat feel, squad presentation, CogSys replay analysis, multiplayer reliability, camera polish, and fixes discovered following the update.
This patch is fully replay-compatible with recordings from 0.1.11.
We've made a substantial presentation pass on squads.
Squad members now behave visually as a small crowd while remaining a single RTS unit underneath.
Members maintain their formation while applying local separation, independent facing, movement variation, and their own movement animations. Once the squad stops, everyone settles cleanly into their assigned positions instead of sliding around trying to continually resolve separation.
Every squad member also now receives their own ground circle and click target, making the actual composition and footprint of a squad much easier to read.
Unit ground and selection circles have been completely rebuilt as lightweight runtime geometry.
The result is much clearer feedback about where units are, where they're facing, and which squad members actually occupy the battlefield.
Combat knockback has been rebuilt into a deterministic impact system.
Impacts now resolve a safe endpoint before presentation begins rather than physically shoving pieces of the unit hierarchy around independently.
This gives us much more reliable knockback while preserving existing movement paths and, importantly, allows the exact same impact result to replicate across multiplayer, Theatre, and Replay playback.
Units and structures can now also author their own Combat Impact profiles, while individual units can have different levels of impact resistance.
The Replay system introduced in 0.1.11 was designed to become more than a way to rewatch matches.
0.1.11b introduces our first durable CogSys Replay Dataset Analyzer.
The analyzer can process .voidreplay datasets and generate structured evidence reports covering:
It also validates replay version and telemetry quality before allowing a recording to be treated as useful training evidence.
Historical recordings, cheat-modified games, and incompatible datasets are explicitly identified rather than silently being mixed into our analysis.
We're already using this process.
This patch contains additional CogSys behavior informed by recorded matches.
Arx Imperium has improved long-term Artifice Lab scaling, better role-specific production priorities, and can now redirect excess Materials workers toward Ichorite when research progression is being blocked.
Arafel can now make a similar strategic economy adjustment when sitting on sufficient Dermis but lacking the Ichorite required for its next research tier.
This is the loop we've been working toward:
Players play. Replays give us evidence. We expand CogSys. Then we test it again.
Pausing the game now provides an additional tactical readability layer.
While paused, visible units receive a lightweight color overlay:
This is entirely local presentation and disappears immediately when the simulation resumes.
The dynamic camera introduced in 0.1.11 has received another polish pass.
Faction cursors now use one unified Cursor Scale setting.
Every contextual cursor scales from the same baseline, giving consistent sizing across attack, resource, edge-scroll, interaction, and normal pointer states.
We've also corrected color and gamma handling for software-scaled cursors so resized faction artwork retains its intended appearance.
Ambient wildlife behavior has received another pass.
The Strand once again begins with up to 50 passive Deernt and Cabra, while the ecosystem's normal spawning system begins replenishing wildlife after the population drops below 30.
Herding has also been reworked with multiple movement lanes, bounded herd sizes, individual destinations, and stronger separation.
Wildlife should now actually travel across The Strand in recognizable groups rather than gradually collapsing into one enormous alien-animal convention.
Research Nodes can now author multiple ordered completion behaviors directly through our Research Behavior system.
We've also completed an audit of the current Arafel and Arx research catalogs, moving 64 runtime research hooks onto explicit authored gameplay gates and outcomes.
For players, this primarily means more reliable research behavior and fewer opportunities for a research effect to become disconnected from what its node says it actually does.
Another multiplayer recovery pass has landed following 0.1.11.
Several dense-match presentation costs have been reduced.
World-space health and construction bars now update their camera projection on a controlled cadence rather than forcing Canvas work every rendered frame while the camera moves.
Malignance doodad visibility checks are also spread across smaller slices instead of periodically processing large batches at once.
We've continued tightening CogSys diagnostics and other late-match systems as well.
0.1.11b is primarily about tightening the enormous amount of new ground covered by 0.1.11 while continuing to put its new systems to work.
Of particular importance to us is the Replay → CogSys pipeline. We now have the foundations for recording matches, extracting useful strategic evidence from them, turning that evidence into new AI behavior, and validating those changes against future games.
That's going to become increasingly important as more players get their hands on the game and inevitably discover strategies we never anticipated.
As always, thank you for playing, testing, breaking things, and sending us feedback.
We'll see you in Atrum.