Thanks for your feedback.
As a matter of fact, I came up with this idea, because I was thinking of the historic development of WW2. Why would anyone fight to the last bullet and the last square meter as Germany did? Assuming that it was not just pure stupidy and/or fanatism, it seems logical, that once you realize that you have no chance of expanding anymore with conventional methods you try to retreat as slowly as possible in order to gain time to develop a "game changer technology". This is what happend historically (the so called Wunderwaffe that was never actually achieved, but the progress was quite significant and a lot of effort was invested)
If this is your strategy you will divert a part of your ressource into this development program, hoping that your scientists will eventually be succesfull. Of course, this means taking chances, but it is a rational approach (from a pure military, not humanitarian perspective).
I calculated the chances of succeeding (based on a target value of 18):
Round 3: 0,5%
Round 4: 16%
Round 5: 50%
Round 6: 79%
Round 7: 94%
Given the high probability of succeeding in Round 6/7 this approach is not simply based on dump luck. So, it could be an interesting option if you want to have a fallback position in case the game does not develop in your favor.
On the other side, it is also a viable option for someone on the superior side with a lot of ressource in order to achieve victory quicker, which also corresponds to the history. You could endlessly battle down your enemy or you could chose to develop such a technology as a shortcut. So, indeed it makes for both sides to think about this option.
Of course, 5pu was only an idea, it could very well be higher or lower. I guess, the only way of finding a sensible cost is by trying. The same applies for the target value.
I based my idea on the Classic WWII or 3rd Edition map. Unfortunately, I am not a programer, so I do not know how this could be implemented for a test.