From Single Moves to Full Convoys

Episode summaryThis episode is an unfiltered deep dive into the hardest remaining system in the rebuild: vehicle pathfinding and convoy behavior. Robert walks through days of live playtesting, debugging, and iteration on the Ajax, Striker, and Mule, showing exactly where vehicles succeed, where they fail, and why that last 10–20% of polish is the most time-consuming part of development.The team breaks down how vehicle AI now evaluates road preference, turn-in-place recovery, reverse maneuvers, obstacle avoidance, and navigation cost layers—and why dense urban environments expose edge cases that simpler RTS maps never face. Convoy movement introduces an entirely new layer of complexity, with formation logic, dynamic spacing, gas usage, and mixed infantry/vehicle groups all fighting each other when systems don’t perfectly agree. Robert openly demonstrates failures: clipping, overcorrection, bad formation anchors, vehicles recoiling off small objects, and rare cases where collision or navigation data breaks down entirely.Alongside vehicles, Michael details ongoing infected model optimization, including atlas-based materials to massively reduce draw calls, enabling far larger battles without CPU bottlenecks. Future polish includes restoring dismemberment chunks, refining VFX, and upgrading animation pipelines to improve fluidity across both operators and RTS units.The episode also previews upcoming combat balance changes—new operator weapons to counter juggernauts, sniper buffs against builders, and experimental heavy canister weapons—while clarifying why cosmetics, infantry skins, and non-critical features are intentionally deprioritized until core systems are rock solid.On the roadmap side, Robert lays out a clear December plan: finish vehicle pathfinding, add remaining vehicle variants, roll in logistics systems like finite resupply, then move into doctrines, airstrikes, and super-heavy units. San Francisco North remains the long-term centerpiece, with island hopping and civilian systems planned once vehicle AI can reliably handle the terrain.The key message is blunt and consistent: vehicles will not ship half-broken. The team is committed to crushing edge cases through aggressive playtesting—even if it takes longer—so convoys work under pressure, not just in ideal conditions.