09-10-2018, 10:32 PM
I've conducted a series of tests and am convinced that the combat resolution tables or equations for the Napoleonic Battles series of games are entirely broken. No modifiers are applied as stated in the documentation and the reported fire efficiency is manipulated by some shadow feature in every instance.
A somewhat simple case in point. 530 man column firing at a line's stated efficiency should be; 530 (men) * 5 (musket range 1 firepower) * 0.33 (column fire penalty) = 874.5 fire efficiency. In the image the engine believes it should be 441 instead. I know that the report dialog's efficiency is actually being used due to the low casualty count. 874.5 x 4 (low combat value) / 1000 = 3.498. The lowest correct possible casualties..
I have found this fuzzy math in every instance of the firing (and modified melee) resolution calculation and am officially crying foul. Another hypothesis I have is due to running these game in a Linux environment they are generating floating point digit errors which are having wild effects. I certainly wouldn't expect JTS to proof their games for Wine compatability. It would be a strange first case of this issue though.
After some scrutiny it appears I was wrong about what the firing values should be. They are apparently halved if a unit was moved that turn. Disregard my naysaying but, I still think the combat dialog is fishy in how it applies most modifiers to the FE value and translates them to casualties.
A somewhat simple case in point. 530 man column firing at a line's stated efficiency should be; 530 (men) * 5 (musket range 1 firepower) * 0.33 (column fire penalty) = 874.5 fire efficiency. In the image the engine believes it should be 441 instead. I know that the report dialog's efficiency is actually being used due to the low casualty count. 874.5 x 4 (low combat value) / 1000 = 3.498. The lowest correct possible casualties..
I have found this fuzzy math in every instance of the firing (and modified melee) resolution calculation and am officially crying foul. Another hypothesis I have is due to running these game in a Linux environment they are generating floating point digit errors which are having wild effects. I certainly wouldn't expect JTS to proof their games for Wine compatability. It would be a strange first case of this issue though.
After some scrutiny it appears I was wrong about what the firing values should be. They are apparently halved if a unit was moved that turn. Disregard my naysaying but, I still think the combat dialog is fishy in how it applies most modifiers to the FE value and translates them to casualties.