I've read probably a billion comments on what to do (if anything) about this issue and its been debated endlessly here at Ironclad as well. Anyways, here it was is changing in the 1.03 patch:
-Carrier's squadron buildrate penalty with enemy units in the gravity well reduced from 25% to 15%.
Now before I'm bombarded with another billion comments, let me explain the situation and why we've made this course correction.
First, our inhouse testing and modelling tells us that 25% is the value we want to balance the flak/lrm/carrier triangle. However, our stats tracking and your feedback suggests it's too harsh. (In terms of stats tracking we can see that the ratio of carriers to lrm over all games played on ICO since the carrier change is not where it should be). This means that we are clearly missing something in our model and in our playstyle. We don't know what it is and it is very frustrating. So what we want to do is reduce the value and see what happens. Based on those results we will either move it back towards 25% or closer to 0%. Expect a patch that does nothing but change this value by 5% in either direction in the near future once we've got more data and feedback. In the meantime we'll be trying to figure out what is missing in our model and playstyle.
Now just to clear up a misconception:
The build rate penalty was not 75%, it was 25%. The value in the gameplay.constants file was -1.0 which means when plugged into the code "1 / (1 - x)" it results in a scalar of .75 which is a 25% reduction. The new value of -.176 will result in a scalar of .85 which is a 15% reduction.
Finally, just so its explicit: we have considered changes to flak, seperate modifiers to fighter/bomber build rate penalties, fighter buffs, lrm nerfs and a lot of other ideas some of which were pretty conservative and some of which are pretty wacky. We do feel this is the best direction after examining all possible alternatives. If this iterative process doesn't give us the results we want we'll open the door again to the other possiblities.