Thanks for the feedback, SackLunch.
My first preference would be to have an option at start up.
That was my first preference too, until Ash correctly corrected my copout thinking here: https://forums.elementalgame.com/329572/get;1934445
In general options like this one would consume an amount of development/testing resources that is disproportionate to the benefit of having them. Or you wind up with one option or the other (or both) being badly imbalanced.
I guess I just don't like the thought that if I specifically pick my techs, why should I be penalized and potentially way behind someone else in techs in the late game?
I don't like that thought either. But you wouldn't necessarily be behind; maybe the opponent uses blind research and just doesn't get the techs they need and you wind up beating them because 75% of your research points is still higher than 100% of theirs or at least yields more actual benefit.
Also, worst to worst, the 75/90/100 (or whatever) numbers should be extremely easy to mod to 100/100/100 and remove the penalty. If you do that you'll be playing a game that hasn't been balance tested, but it shouldn't be much of a problem.
If I research laser I & II, there is no penalty, but if I go straight to laser III or above, there becomes a progressively higher penalty if I'm not researching other techs.
This would be neat, but it would be more complicated from an implementation standpoint whereas I think the earlier proposal should be pretty simple on the programmers (except the AI guys, that could be tricky). A big consideration for a feature of this kind is to minimize impact on development resources and thus not making it very complicated.
For the sake of argument, I imagine what you would propose would still use the tech-category idea, and that the "get-penalty-for-researching-tech-in-category-X" algorithm would look like this:
- for each category
-- get sum of "tech points" researched in this category
- if category X's tech point total is more than 30% of the tech point grand total
-- reduce efficiency to 90%
- else if category X's tech point total is more than 40% of the tech point grand total
-- reduce efficiency to 80%
So it might not be so bad to program but I think a bit worse than the blind/category/specific idea.
The other problem is that it wouldn't really move away from the rigid "build order" approach to techs where you figure out the optimal path through the tech tree to optimize for a certain strategy. It just makes the computation of that path more complex by the need to figure out how to shuffle the order of the different category techs to minimize the efficiency loss.
But perhaps people will like your idea better, or refine it into something better. Anyone?