How many scientists per Projekt ?

[Zack]

Should i stop assign more scientists to a project if the progress is shown as excellent ?

[Max_Caine]

You can assign as many as you like. Each scientist you add to a project does less work than the last one, so there is a point where adding more scientists isn't worth it. However, the exact formula is secret.

[KevinHann]

I think there is a limit, I once hit it during 21 Stable, maybe around 100 scientists. Haven't ever built that many labs since though and it's a bit hazy now.

[Jean-Luc]

I always put only as many are needed for excellent progress and then spread them out, you get the most value out of your scientists that way I feel. You can actually micro manage your research. As time goes on less and less guys are needed for excellent progress so you can gradually move scientists from one project to another.

[Sandyxx]

i'm think, that shown status (good-average-excellent, etc.) corresponds to how much time is needed to finish the research and not to it's status.

each research has human-day value. so in the end if we add scientists to the excellent research will be completed in, lets say 2 days. but if we hack the game and assign 99 scientists to research it will be completed much faster, maybe in matter of hours.

so imo best way to research something valuable is to assign maximum scientists to that research, instead of spreading research potential.

because if we will spread researches (e.g. armor and laser weapons)

lets say, armor needs 120 human-days and lasers need 60: we got 30 scientists to assign.

if we assign them 20 to 10, then we will get both researches done in six days.

if we assign them 0 to armor and 30 to weapons, then weapons will be finished on second day. and after that we will research armour in 4 days.

so in the end, we will obtain both researches on sixth day, but in second case, we will already get laser weapons and we can produce them. so we minimize state when we doesn't have anything.

[bills6693]

You can assign as many as you like. Each scientist you add to a project does less work than the last one, so there is a point where adding more scientists isn't worth it. However, the exact formula is secret.

Scientists have diminishing returns? Did not know that, thats awesome to know. Now I actually have a reason to spread them out, before this I thought it was just pointless.

[Slandebande]

i'm think, that shown status (good-average-excellent, etc.) corresponds to how much time is needed to finish the research and not to it's status.

each research has human-day value. so in the end if we add scientists to the excellent research will be completed in, lets say 2 days. but if we hack the game and assign 99 scientists to research it will be completed much faster, maybe in matter of hours.

so imo best way to research something valuable is to assign maximum scientists to that research, instead of spreading research potential.

because if we will spread researches (e.g. armor and laser weapons)

lets say, armor needs 120 human-days and lasers need 60: we got 30 scientists to assign.

if we assign them 20 to 10, then we will get both researches done in six days.

if we assign them 0 to armor and 30 to weapons, then weapons will be finished on second day. and after that we will research armour in 4 days.

so in the end, we will obtain both researches on sixth day, but in second case, we will already get laser weapons and we can produce them. so we minimize state when we doesn't have anything.

I agree, I have used the "progress" tab for estimating the time needed for the project to complete. I also agree with your second point, that if the two things being researched (and you aren't manufacturing anything super important) it would be optimal to have the two researches completed with at least some time to start manufacturing the first item.

One thing though, to the more knowledeable people around here:

Does the diminishing returns of scientists apply on a per project basis? I had always thought it was applied on a per base basis. Or maybe it is a combination? That would at least make the most sense to me, as they would all be sharing the same lab whether they are working on different projects or not, but the "too many cooks spoil the broth" saying would also be applicable in most cases, as there is a point where throwing additional hands at a problem is not the solution.

[Sandyxx]

ok here is info from test:

Researches.HunterTank 80

Researches.MIGInterceptor 80

started at sep 01 4:58 ~ 5:00

1) both only excellent (21 scientist each) : both researches are finished at sep 05 10:35

4 days and 5,5 hours = 101,5 hours *21 = 2131.5 human-hours for each. total 4263

2) 40 scientists set to each research: both finished at sep 03 16:40

2 days and 11,6 hours (11 hours 40 minutes)= 59.6 hours *40 = 2384 human-hours for each. total 4768

3) 80 scientists to mig and after that 80 to hunter

mig is finished at sep 02 20:40

39.6 *80 = 3168 human-hour

hunter is finished at sep 04 12:21

79,3 *80 = 6344 human-hour total

so in the end, i was wrong, and with excessive amount of scientists some work is wasted

Edited June 19, 2014 by Sandyxx

[Zack]

Nice

[Max_Caine]

It might be worthwhile working out what the exact level of diminishment is, because then you can work out the most efficent spread of scientists.

[Dranak]

I think I remember it being 1% per scientist, but I'm too lazy to math.

[Yianyan]

The number is a secret? From the manual: Scientists assigned to a project suffer from diminishing returns – the first scientist works at 100% efficiency, the second at 99% efficiency, the third at 98% and so on.

Is this a lie or did everyone just miss it?

[Max_Caine]

Shows you how often I read the manual. If that's what the manual says it is, then that's what it is!

[bills6693]

Yes, and therefore the 'ideal' spread of scientists, to get the most work done, is to be equal between all possible projects. However then you'll get nothing fast. Thus I'd recommend, personally, a balance (and that is what is intended, I imagine).

I would say don't assign more than about half of your scientists to one project, and assign your other ones split between other projects. That way you can get things done that you need, quickly - but you are not wasting too many man-hours rushing that tech, and are getting better progress on other techs instead.