Station and ship rates are measured per hull on two station hulls and two ship hulls with holds that
sixteen gear picks fill, on a station whose first bay refuses, and on the shipped catalog's own hulls.
The ship band is 3000 seeds so chances of .12, .15 and .05 fall outside it. These tests fail on the
generator as it stands (consumable stocked after the gear); the next commit fixes the generator.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A mutant that changed a constant moved the test's expectation with it and survived (station chance 1,
ship chance 0 and 1, station units 3).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The gated-versus-plain comparison alone cannot see an unconditional draw (both catalogs draw).
Values measured on origin/master d4a27def with the same fixture. The roll-before-availability
mutant survived the full suite without this pin.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
LoadoutGenerator.AvailableProducts and RandomProducts now range over CraftedItemData; only
equippable gear needs a hull home. StockConsumables is the last step of GenerateShipLoadout and
GenerateStationLoadout: it draws nothing when no consumable is on offer. Unbuilt/untested at
this commit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>