Japanese AI Conversions & Upgrade Chains — Has Anyone Optimized These for AI‑vs‑Human Play?
Posted: Wed Mar 25, 2026 10:34 pm
I’m doing a ground‑up rebuild of my old scenario and I’ve reached the point where I need to take a hard look at the Japanese ship conversions and upgrade chains. As most of us know, the Japanese AI never uses optional conversions — it only follows upgrade paths. That means a lot of historically important IJN conversions (PBs, APDs, AKVs, DMS, etc.) never happen unless they’re modeled as upgrades instead of optional conversions.
Before I start restructuring these chains, I wanted to ask the community:
Has anyone already worked on optimizing Japanese conversions specifically for AI‑vs‑Human scenarios?
I’m not looking for files — just your experience, notes, or general approach. In particular:
- Have you forced certain conversions into the upgrade chain so the AI actually uses them?
- Did you pre‑convert any ships in the OOB to give the AI a more realistic force mix?
- Did you restructure the upgrade trees to simulate conversions the AI would never choose on its own?
- Are there any “best practices” you’ve found for improving Japanese AI behavior through conversions?
- Did you identify any conversions that shouldn’t be forced because they weaken the AI?
My goal is to give the Japanese AI a more historically accurate and functional fleet without giving it anything ahistorical or unbalanced. If anyone has already explored this problem — even partially — I’d appreciate hearing what you tried and what worked (or didn’t).
Thanks in advance to anyone willing to share their thoughts or experience.
Before I start restructuring these chains, I wanted to ask the community:
Has anyone already worked on optimizing Japanese conversions specifically for AI‑vs‑Human scenarios?
I’m not looking for files — just your experience, notes, or general approach. In particular:
- Have you forced certain conversions into the upgrade chain so the AI actually uses them?
- Did you pre‑convert any ships in the OOB to give the AI a more realistic force mix?
- Did you restructure the upgrade trees to simulate conversions the AI would never choose on its own?
- Are there any “best practices” you’ve found for improving Japanese AI behavior through conversions?
- Did you identify any conversions that shouldn’t be forced because they weaken the AI?
My goal is to give the Japanese AI a more historically accurate and functional fleet without giving it anything ahistorical or unbalanced. If anyone has already explored this problem — even partially — I’d appreciate hearing what you tried and what worked (or didn’t).
Thanks in advance to anyone willing to share their thoughts or experience.