Then I think about it, the tech reduction get stronger and stronger the larger you get it.
Lets take a technology that have 100 day research time and that tech is researched over and over. You research the -2% tech and thus have a tech reduction of -2%.
Then each 98 days you are saving 2 days for about 7.5 saved days per year.
Now lets say you have a tech reduction of -50% and research the -2% tech.
Then each 48 days you are saving 2 days for about 15.2 saved days per year.
So what Im saying is that you want to stack the research reductions.
Now a realistic tech reduction before -2% tech is between -2% to -10% depending on which trade law you use and if you use companies it is possible for some technologies to have as much as -20% tech reduction before the -2%.
If you have -10% you are saving 2 days every 88 days and if you have -20% you are saving 2 days every 78 days.
Then you get the -3%. Then you will be saving 5 days each 85 days with -10% and saving 5 days each 75 days with -20%.
You are thus undervaluing these techs. With -10% reduction the payback time is 2.71 years for -2% research reduction and 2.62 years for -3% research reduction if you have 4 research slots. The -3% also improve the -2% so the payback time may be even shorter then I have listed.
Then there is the fact that you can not develop encryption and decryption without the computer techs, how valuable these are I don't know but it is something to consider.