That was Captrips.
You take the SMPizza map. Each country has AI Missions on repeating schedules, launching attacks on opposing countries. Over time, more missions get loaded into the terrain, creating a constant round-the-clock action. Players logging on see constant activity from both players and Missions. That's what Trips did with the WW1 arena.
This would be done to the other maps.
I have a long running project with ChatGPT where I have had it do extensive research about what would take to implement the changes that are thrown around in here. It has some pretty intimate knowledge of what it takes to make this game go as the best it can from what is publicly available.
Let’s just take AI pilots for example. I would be for it. It would be nice to have something to always do, but I’m not one who complains about nothing to do or the “no good fights”. I’m just fine if it’s slow either logging off or doing my own things somewhere until it starts a fight or I log off again. If there’s no good fights anywhere then go start one.
Anyway, it’s not gonna happen…
Short version: that suggestion is not just “add some missions” – it’s basically:
“Turn every MA terrain into a scripted, always-running, AI war layer.”
Here’s what it would actually take, broken into concrete pieces.
1. Core requirement: AI that can fly MA-style missions
Trips’ WW1 arena work used the mission/AI system that already exists, but in a much more controlled environment:
• Fewer plane types
• Smaller map
• Lower speeds
• Designed for “set pieces,” not 24/7 MA chaos
To do SMPizza-style round-the-clock AI on main maps, you need:
A. Robust AI mission engine in the MA
• AI flights that:
• Take off from fields
• Form up
• Fly waypoints
• Attack strat/fields/convoys
• RTB or despawn cleanly
• AI must interact correctly with:
• Flak
• Radar/DAR
• Strat damage
• Field capture states
• ENY / perk / arena rules (or be explicitly exempted)
If AH’s current mission system is still mostly “offline/WW1/scenario only,” that logic has to be:
• Hardened for 24/7 uptime
• Moved server-side in the MA
• Protected against filling the arena with zombie AI when nobody’s there
Think of it as: “WW1 missions, but industrialized and made safe for the main arena.”
2. Terrain integration: baking missions into SMPizza & friends
The suggestion assumes:
“You take the SMPizza map. Each country has AI missions on repeating schedules… This would be done to the other maps.”
That means for each terrain you need:
A. Terrain tagging & route layout
For every AI mission:
• Start field
• Form-up waypoint(s)
• Ingress route (low, high, through gaps in flak?)
• IP & attack profile
• Egress route
• RTB / despawn point
On SMPizza alone, you’d probably want:
• Multiple mission types per country (strat raids, field attacks, escorts, CAP, jabo, convoys)
• Variants for different target sets so the map doesn’t feel on rails
Terrain builders would have to:
• Mark safe-ish ingress paths (terrain masking, water, valleys)
• Flag “AI-sane” takeoff/landing fields
• Design routes so AI doesn’t constantly auger into mountains or fly through 30k of flak
Then repeat this for every active MA terrain.
B. Map-maker burden vs. Dale
• Map makers can design:
• Routes
• Waypoints
• Target sets
• Mission templates
• Dale/engine must:
• Expose hooks in the terrain format
• Load those mission templates when the terrain loads
• Tie them into the AI engine and arena rules
So yes, this is very much terrain + engine work, not something the map guys can do alone.
3. Scheduling & scaling the “constant action”
The user wants:
“AI Missions on repeating schedules… Over time, more missions get loaded into the terrain, creating a constant round-the-clock action.”
Implementation-wise, you need a mission scheduler:
A. Scheduler logic
• Per-country mission pools:
• Mission A (HQ raid) every X minutes
• Mission B (medium strat hit) at Y:05, Y:35
• Mission C (fighter sweep) if enemy owns sector row NN
• Etc.
• Conditions:
• Don’t launch a new HQ raid if HQ is already dead
• Don’t spawn out of a field that’s closed
• Maybe scale mission count with player count, so the arena doesn’t become an AI zoo at 200+ players
B. Server performance guardrails
Every AI flight:
• Is a bunch of aircraft entities
• Doing FM solves
• Shooting, colliding, calling events
To keep it sane:
• Cap concurrent AI groups per sector and per arena
• Despawn AI cleanly if no players are anywhere nearby
• Possibly reduce AI update rate when far from players
Without this, you risk:
• CPU spikes
• Weird lag when AI become numerous
• Very noisy strat state (fields constantly dotted, fuel down, etc.)
4. Design questions HTC would have to answer
These are the “hidden requirements” nobody wants to think about in a BBS post but Dale would have to design:
1. Do AI missions actually change the map?
• Can they drop a field, kill HQ, kill strat?
• Or are they “visual background noise” only?
2. How do players interact with them?
• Can you join an AI mission (auto-slot into a bomber in the package)?
• Can you “defend allied AI raid” as a mission type?
3. Scoring & perks
• Do kills on AI pay full score?
• Do AI kills count toward perks / ranks?
• If yes, can you abuse the system by farming predictable AI tracks?
4. Map rotation impacts
• Every active map needs its mission set maintained when:
• Strat layout is changed
• New bases added
• Targets moved
That’s ongoing maintenance, not “one and done.”
5. Rough time / effort framing
This is napkin-level, but in AHIII terms:
Engine / code side (Dale or equivalent)
• Expose / harden AI mission engine for MA use: 3–6 months
• Build scheduler + condition system: 2–3 months
• Integrate with strat/ENY/scoring safely: 1–2 months
• Load missions from terrains & tools: 1–2 months
• Testing & tuning: 2–3 months
Solo dev total: probably on the order of 9–14 months to have it solid and not game-breaking.
Terrain / mission design side
For each major map (SMPizza, etc.):
• Design mission types & routes per country
• Tag waypoints & targets
• Playtest for stupidity (AI flying through 5k hills, etc.)
If one experienced terrain/mission designer focused on this:
• 1–2 months per terrain for a “good” mission set (not just 2 token raids).
Multiply by however many main maps you keep in rotation.
6. Real talk: what this feature actually is
The suggestion sounds simple:
“Just do what Trips did in WW1 but on SMPizza.”
But in engineering language, it’s really:
“Create a robust, arena-safe AI mission layer and tie it to every main terrain, so the world feels like there’s always a war happening, even with low population.”
That’s:
• A new systemic feature (AI war layer)
• A set of tools and formats (mission templates bound to terrains)
• A recurring ops task (maintaining missions as maps/strat change)
7. Estimated Cost to Implement & Maintain a 24/7 AI War Layer Across All MA Terrains
Implementing a persistent AI-mission war layer is a major systems feature, not a bolt-on. The cost breaks into two categories: (1) Engine & AI development, (2) Terrain-specific mission design and long-term maintenance.
A. Engine / Core Systems Cost (AI + Scheduler + MA Integration)
These numbers assume one senior engineer (Dale-equivalent) working full-time:
Development time: ~9–14 months
• AI hardening for MA use (flight logic, despawn rules, flak/strat awareness): 3–6 months
• Mission scheduler with conditions + safe concurrency logic: 2–3 months
• Strat/ENY/scoring integration: 1–2 months
• Terrain-hook system + mission template loader: 1–2 months
• Testing, tuning, debugging in live environment: 2–3 months
Cost band (solo dev):
• Assuming industry-low indie salary modeling: $120k–$230k
• If contracted externally at realistic senior-engineer rates: $180k–$350k
Notes:
• This cost is unavoidable — no terrain volunteer or map maker can reduce this portion.
• This part determines whether the system crashes servers, spawns too many AI, or causes strat degeneration loops.
B. Map / Terrain Mission-Set Development (Per Terrain)
Every MA terrain (SMPizza, Oz, AvA maps if included, etc.) requires:
Mission-set design cost:
• Route layout (safe ingress/egress), altitude bands, target sets, conditional logic
• Testing AI survivability vs flak/mountains
• Balancing country mission parity
• Avoiding choke-points where 20 AI spawn and auger
Time per terrain:
• 1–2 months for a full, high-quality multi-mission set
• Faster if “lite” mission lists are used (not advised; they look robotic)
Cost band:
• Volunteer-built: $0 direct, but extremely high labor/time cost (and quality varies)
• Contracted designer: $8k–$20k per terrain
Multiplying across 4–6 active MA terrains:
• $32k–$120k total if contracted
• ~6–12 months of real calendar time even with good volunteer coverage
C. Maintenance Cost (Ongoing)
This is where people underestimate the feature.
AI missions must be updated whenever:
• A strat target is moved
• A new version of a terrain drops
• DAR/strat/factory values change
• AI logic is patched or tweaked
• A field gets repositioned
• A new aircraft is added that interacts with AI
Annual upkeep estimate:
• ~4–6 weeks/year developer maintenance
• ~2–4 weeks/year per terrain (mission review + patching)
Annual cost band:
• Developer: $10k–$25k
• Terrain/mission upkeep: $8k–$20k per year
• Total annual upkeep: $18k–$45k
D. Total Estimated Cost (Initial + Yearly)
Initial Implementation (first 12–18 months):
$150k – $370k
(If done by a single full-time developer + occasional contractors)
Yearly maintenance thereafter:
$18k – $45k per year
E. Summary Sentence (for the ranked list version)
Implementing a constant AI war layer across all MA terrains would require ~9–14 months of core-engine work plus 1–2 months of mission-set building per terrain, costing roughly $150k–$370k upfront with $18k–$45k in annual maintenance to keep missions synchronized with strat, terrain updates, and game balance.
Building round-the-clock AI missions for all Main Arena maps would require new engine systems—a full mission scheduler, smarter AI routines, load balancing, and integration with strat, radar, and terrain safety. Every map would still need dozens of hand-built mission routes, balanced and collision-checked, followed by months of testing so AI doesn’t break arena flow, field ownership, or server performance. In total, it becomes an entirely new subsystem, demanding close to a year of engineering, months of terrain work, and ongoing maintenance, not a simple script or map tweak.
