[Bug Report] Cloak and warp core stabilizer scan-resolution stacking differs from game #2805

Bug Report

Expected behavior:

A cloaking device and a warp core stabilizer fitted to the same ship appear to share scan-resolution stacking penalties in the live game. With the minimal Nereus fit below, the game fitting window displayed 106 mm (neither module), 53 mm (cloak only), 59 mm (stabilizer only), and 32 mm (both). These are integer UI readings, not precise server-side values.

The in-game pilot does not have all skills at V. This report does not equate its exact base value with an all-V pyfa pilot. Instead, the within-pilot readings rule out simply multiplying the two individual penalties: C * W / B is about 29.5 mm, and display rounding cannot explain 32 mm.

Actual behavior:

In a minimal pyfa EOS reproduction with all skills V, modules online and inactive, no implants or other modules, pyfa reports 106.25 / 53.125 / 59.5 / 29.75 mm for the same four states. Both modules are present and applied, but the penalties do not reduce one another. This is a pyfa core result, not a comparison-script state mismatch.

Detailed steps to reproduce:

1. Import the EFT fit below into pyfa and use an all-V character for a reproducible pyfa baseline.
2. Keep both modules online and the cloak inactive. Read scan resolution.
3. Remove each module in turn to read the two single-module values and the empty-hull value.
4. Compare the combined-to-single relationship with the live fitting-window readings above.

Fits involved in EFT format (Edit > To Clipboard > EFT):

[Nereus, Cloak and WCS scan resolution]

'Stoic' Core Equalizer I

Prototype Cloaking Device I

Release or development git branch? Please note the release version or commit hash:

Reproduced from source commit 8b04f3b (2026-08-01), pyfa database build 3424810. The current master source inspected on 2026-09-30 still has the same effect grouping; I have not run the current release GUI as a separate reproduction.

Operating system and version (eg: Windows 10, OS X 10.9, OS X 10.11, Ubuntu 16.10):

Windows 11. The numerical reproduction used pyfa's EFT importer and EOS fitting core.

Other relevant information:

A likely source of the difference is that Effect854 puts the cloak scan-resolution multiplier in penaltyGroup='cloakingScanResolutionMultiplier', while Effect2645 applies the stabilizer multiplier in the default group. ModifiedAttributeDict penalizes each group independently. As a result, the two multipliers are both treated as first-in-group and multiply directly.

Could a maintainer verify whether this separate-group rule still matches the live game? The integer game UI readings establish a discrepancy for this module pair, but do not prove the exact stacking coefficient or other cloak/WCS combinations. No in-game screenshots are available for this report.