i have thought about it in the past, and at some point i assumed that there are, say, 5 puzzles related to one research topic, that each researching scientist has an independent probability of solving (no collaboration), and did a little calculation.
if i remember correctly, the resulting formula was of the type
researchSpeed = numberOfScientists / (const + numberOfScientsis).
a fairly common formula in chemistry, physics, biology, etc.
i do not remember how i derived it from the probabilities, though.
say, if const=50, then
10 scientists at the same project would work at the efficiency of 8.5 scientists each on separate project,
50 would work liike 25.5
and 100 would work like 34.
[...]If it could be implemented into the research system by adding an attribute to each research item like Diminishing returns (nbr of scientists) after which the effectiveness of each scientist beyond the limit diminishes by some fairly brutal formula. Naturally if you kept the number hidden, it would add an element of uncertainty (at least up until people go poke inside the code[...]
i would not suggest to keep the number hidden, because then there would be privileged people who have read the spoilers.
i would rather suggest to display the resulting efficiency of the research as a number in the gui.
[...]I find myself often assigning scientists to up to three projects simultaneously just to get a headstart on something while not needing the higher priority thing quite that urgently.[...]
when you give it some thought, you might as well assign them all to same project and get to use the results of that project while you research the next project : )