Supported launch scope
Ship Damage Trainer
The trainer displays the current model availability alongside the catalog. Availability is not a claim that every matchup has been validated. Model delivery can also fail; unavailable geometry does not produce aim recommendations.
Aiming estimates cover main-battery shots against available battleship, cruiser and destroyer models. Destroyers have no citadel target. Other classes are inspection-only. Each entry uses one recorded hull and base statistics. AP advice requires matching range-dependent shell data; unresolved armor remains gray. Green indicates a preferred estimated target, never a guaranteed hit.
Armor exports contain thickness intervals. Layered penetration, ricochet, overmatch and sampled fuze timing are approximations. Water-entry shots, dispersion, damage saturation, module damage, commander skills and equipment effects are not fully reproduced.
Recorded replay
The included Clan Battle sample is available for playback. Signed-in members can submit supported Clan Battle replays. Imports depend on decoder compatibility and published map and ship assets; unsupported uploads show an error without consuming replay allowance. Only observed positions and events are reconstructed; unseen enemy movement, missing ribbons and achievements cannot be recovered. Shell flight paths are approximations. The viewer uses exported Unreal assets in a browser; it is not a live Unreal Engine session.
Experimental battle simulation
The preview supports equal-tier V–X surface ships, 1–7 per team, across seven listed maps. It uses seeded, rule-based decisions, approximate combat and simplified terrain. Historical and approximate map layouts are labeled. Aircraft, submarine combat, detailed armor ballistics, saturation and commander/equipment effects are unsupported. Results cannot establish a team's win probability or predict human play.
Validation coverage
Representative tests cover Bismarck, Montana and Shimakaze firing at Tallinn, Des Moines, Yamato and Shimakaze at 6, 12 and 20 km and five target aspects. Tests check internal consistency and conservative armor recommendations. They are not a comparison against instrumented live-game damage measurements. Playback checks cover chunk loading, cancellation, deterministic simulation and movement bounds; visual coverage varies by available model and replay assets.