I was curious to see if different postures/ROE would affect the chances of successfully destroying railroads. It does with the amount of power in the unit, as described in this post. ("A check is made of D100 <{less than} Stack Strength. +25 to stack strength if attribute “*pillage*” is applied to model".) Based on the following experiment, it might also matter what the posture/ROE is as well, but I need help from someone more statistically-minded than I to confirm.
In the Poland scenario I sent five units to five different regions with railroad: Skierniewice, Deblin, Lubartow, Biala Podlaska, and Kobrin. Each unit had some unique feature. I then set them all to destroy railroad and advanced the turn to Late February. Afterwards I set the turn back Early February and tried again. After 20 trials I totaled everything up. Here are the details/results:
Region ... Unit ... Power ... Posture/ROE ... # of Successes ... % Success
Skierniewice ... 1 cav ... 17 ... Defend/Retreat if Engaged (Def/Ret) ... 3 ... 15%
Deblin ... 2 cav ... 32... Def/Ret ... 6 ... 30%
Lubartow ... 1 cav ... 17... Passive ... 6 ... 30%
Biala Podlaska ... 1 cav ... 17... Offense/Retreat if Engaged ... 17 ... 1 ... 5%
Kobrin ... 2 elite cav w/ 1 horse arty ... 85... Def/Ret ... 16 ... 80%
With Skierniewice, Deblin, and Kobrin the results followed the amount of power perfectly, just as the post linked above predicted. But it also looks like posture/ROE affects success as well. Lubartow was on passive, but did twice as well as its power would have us predict. But the real surprise was Biala Podlaska, set to Offensive. Only one out of twenty tries resulted in success. Perhaps the troops were too busy looking for someone to shoot at to bother with busting up some railroad?
Maybe twenty trials like this isn't enough to justify these conclusions, but it would seem that raiders should avoid offensive posture if railroad breaking is their goal. Probably they wouldn't be set that way anyway, but maybe they should be set to passive instead?