Most effective research method

[Micro102 (OP)]

From watching the way the research points work, I've come to think that the best way to research is to assign a project to every scientist you have, even if you have 30 of them, and then give the extra labs to the ones with the longest research time. This could lead to several extremely long research periods, but in the end it is faster then pouring everything into a couple of researches at a time.


Can someone confirm this?

[Father Tim]

Nope, although it depends on what you mean by 'best'.  If you have ten labs, two equal-RP projects, and two scientists (each with 20% bonus and the ability to control ten labs) it will take the same total amount of time to research both projects however you split up the labs, but if you do all ten labs on one 'til it's finished then all ten on the other, for half the time you'll have one project already finished.  That may not make much difference if your two projects are BFC Speed and BFC Range, but if they're a pair of missiles you can be manufacturing one that much sooner, and if they're increased mining & increased production, that's a lot of output you'd be missing out on.

Which means it's highly dependent on what you're researching, how many scientists you have vs how many labs, whether their bonuses are in the right field, and how many small projects you have lined up.  And it changes everytime you build a new lab, recruit a new scientist, or a scientist increases her abilities.

[sloanjh]

From watching the way the research points work, I've come to think that the best way to research is to assign a project to every scientist you have, even if you have 30 of them, and then give the extra labs to the ones with the longest research time. This could lead to several extremely long research periods, but in the end it is faster then pouring everything into a couple of researches at a time.


Can someone confirm this?

Here's my take: viewtopic.php?f=19&t=2102&p=20516&hilit=training+labs#p20516

If researchers were immortal, and all trained at the same rate (rather than randomly), then you would only want the best researcher in each field working non-stop.  Since they aren't immortal, you want to keep 2-3 trained up in case your best one dies.  Since training advancement is random, it also makes sense to train all of your researchers that are near the top in case one gets lucky and pulls ahead.  It doesn't make sense to train a low-skilled one if you've already got several high-skilled ones in the same field.

John

[Micro102 (OP)]

Now hold on, if you have 2 50 year research projects, and add an extra lab to one of them, it could go down by like, 10 years, add another and then it goes down by 9, then 8 then 7 then 6 until it only goes down by a couple of months whenever you add a new research lab.

So wouldn't it be better, if you only had 4 labs, to put one on each of them, removing 10 years form each, leading to 20 years shorter, instead of putting 2 of them on 1 research project, and only removing a total of 19 years?

[sloanjh]

Now hold on, if you have 2 50 year research projects, and add an extra lab to one of them, it could go down by like, 10 years, add another and then it goes down by 9, then 8 then 7 then 6 until it only goes down by a couple of months whenever you add a new research lab.

So wouldn't it be better, if you only had 4 labs, to put one on each of them, removing 10 years form each, leading to 20 years shorter, instead of putting 2 of them on 1 research project, and only removing a total of 19 years?

Who are you replying to, Father Tim or me?

Father Tim's analysis is spot on.  I think the issue you might be having is thinking in years rather than research points.  There's no such thing as a "50 year research project"; there is such a thing as a "50,000 research point project".  If you've got two of them, then you have to do 100,000 points of research to get them both done no matter how you slice it.  If each lab can do 1,000 points per year and you've got 4 labs, that's going to take you 25 years (25 years * 4 labs * 1,000 points/(lab*year) = 100,000 points).  As Father Tim says, the only choice you've got is if you work only on the first one (which is done at t=12.5 years) then only on the second (which is done at t=25 years), or if you work on them both equally, in which case neither is done until t=25 years.

John

[Micro102 (OP)]

Hmmm, thinking it over, focusing on the times instead of the points was probably the wrong idea. (and i thought i was smart spamming 10 research to be done in 5 years   )

So, putting all research labs into one project is the best way to go?

[sloanjh]

Hmmm, thinking it over, focusing on the times instead of the points was probably the wrong idea. (and i thought i was smart spamming 10 research to be done in 5 years   )

So, putting all research labs into one project is the best way to go?

Read the link I posted upthread.

John

[Micro102 (OP)]

Ah, now i get it, thx.