Is the research "Lab Research Speed" necessary?

[LazyGao]

Hi guys

Maybe its already a common understanding on this topic, if the conclusion i made is correct. But maybe i'm wrong in some step, so here is the thinking:

I've been thinking about the research item "Lab Research Speed", and it just came to me that if i put all the investment of the research into real things: research labs themselves, how many of will i get?
So i tried this calculation(only in iron and copper):
(not good at composing tables)
1. Research Labs:
every lab costs 70 irons and 50 coppers, no more than 4 belts(sushi belt, belt per lab) which is 11.5 iron(red belt) * 4 = 46 irons, 1 fast inserter is 8 irons and 4.5 coppers, neglecting electric poles.
so totally 124 irons and 54.5 coppers.

2. Research levels:
Lab Research Speed lvl 1
100 red + 100 green science packs = 750 irons + 250 coppers = approx. 6 labs(iron-wise) or 4.5 labs(copper-wise) = 4.5 labs at least.
the research increases 20% performance so 4.5 / (20% / 100%) = 22.5,
so if i already have more than 22.5 labs, the research itself is better than building more labs.(not even saving irons, only copper-wise)

Lab Research Speed lvl 2
200 red + 200 green science packs = 1500 irons + 500 coppers = approx. 12 labs(iron-wise) or 9 labs(copper-wise) = 9 labs at least.
the research increases 30% performance so 9 / (30% / 120%) = 36,
so if i already have more than 36 labs, the research itself is better than building more labs.(not even saving irons, only copper-wise)

it is not bad so far, because normally i build 40~50 labs in all.

Lab Research Speed lvl 3
250 red + 250 green + 250 blue science packs = 10375 irons + 3000 coppers = approx. 83 labs(iron-wise) or 55 labs(copper-wise) = 55 labs at least.
the research increases 40% performance so 55 / (40% / 150%) = 206.25,
so if i already have more than 206.25 labs, the research itself is better than building more labs.(not even saving irons, only copper-wise)
this number is already not quite reasonable.

Lab Research Speed lvl 4
500 red + 500 green + 500 blue science packs = 20750 irons + 6000 coppers = approx. 167 labs(iron-wise) or 110 labs(copper-wise) = 110 labs at least.
the research increases 50% performance so 110 / (50% / 190%) = 418,
so if i already have more than 418 labs, the research itself is better than building more labs.(not even saving irons, only copper-wise)

Lab Research Speed lvl 5
500 red + 500 green + 500 blue + 500 purple science packs = 44000 irons + 14150 coppers = approx. 354 labs(iron-wise) or 259 labs(copper-wise) = 259 labs at least.
the research increases 50% performance so 259 / (50% / 240%) = 1243,
so if i already have more than 1243 labs, the research itself is better than building more labs.(not even saving irons, only copper-wise)

save for the gold science packs... insane numbers

so my conclusion is that, normally if you make 40~50 labs, speed research level 1 and 2 are reasonable(again, lvl 1 saves you both on irons and coppers, lvl 2 is only good for coppers)
the levels 3-6 is not quite a good deal, just build more labs will perform better than the research lvls.

sure i didn't calculate the electricity usage of more labs, but i don't think that will change the result much.

[leitk]

Now add 2 productivity 3 modules to each lab (so you need fewer beakers). And speed beacons around them (so you need fewer modules in total).

[DaveMcW]

Yes, the real cost of a lab includes 2 productivity modules. You really want to use productivity for higher level research.

[mrvn]

One other thing to consider:

You can build more science labs while researching else. But you can't research something else while researching "Lab Research Speed".

So you not only have to make up the cost of science labs but also the time lost.

[BlakeMW]

It's fairly well known I think that the non-green science research speed upgrades are not very good value.

The blue science ones might be worthwhile if you're using prod3 modules, but it might well be cheaper to just add some speed3 beacons.

[Jap2.0]

Thanks for doing the math. Also note that one inserter and 4 belts is a very minimum there are other things to think about as well - what if you are using blue belts? You don't necessarily need belts, you can just use requester chests leading into labs inserting into each other - but then you will need lots of requester chests or inserters. What if you don't do a sushi belt (I know many people don't), but instead do half a belt of each science? This increases the cost to 12 belts and 4 inserters per lab. You also have to consider the extrra space used by new labs, the power costs (of building new labs? of researching higher speeds? does an increased spped from research increase power consumption?), and plenty of other things. Of course, that doesn't mean that it's a good research.

[5thHorseman]

For me, so long as I keep the (automated) research going on at most any speed, I have far more things researched than I'm doing. Other than the first few techs at the very start, and Advanced Oil Processing because you want it ASAP after getting oil running, I never, ever need something I've not researched hours before.

That said, I always research the "Research speed" nodes, for no better reason than because they're there.

[Hannu]

It is true that there is no reason to give a priority for lab research speed. But at least if you have normal settings and reasonably sized production there will be time you can not build new production as fast as you produce research. It begins before advanced oil production. Then there is some "necessary" blue science things but after that there is some time for less important research before next tiers of science. At least I choose then many research things I do not need any time soon or perhaps ever and it is nice to take research speed then.

Someone let their factory to stop in that situation to avoid pollution and biter attacks, but I like to see everything running and it also helps to find and fix possible bottlenecks early. I have to build defenses against biters but it gives even more activity for me and my production facilities.

[MeduSalem]

For me, so long as I keep the (automated) research going on at most any speed, I have far more things researched than I'm doing. Other than the first few techs at the very start, and Advanced Oil Processing because you want it ASAP after getting oil running, I never, ever need something I've not researched hours before.

Same here... I'm almost always far ahead in research than what I'm currently doing in my factory (building stuff takes always more time than expected), except for Advanced Oil Processing which is always somewhat a stressful/frustrating experience which I'm not really looking forward to and which I'm happy when overcome.

That said, I always research the "Research speed" nodes, for no better reason than because they're there.

Me too.

Mostly I'm researching what's currently the cheapest to research (except if I need something desperately) so things take gradually longer and become gradually more complicated and when a new science pack gets added to the mix I usually expand the labs and entire infrastructure anyways to keep up... Eventually lab research speed becomes the cheapest research and that's when I knock it of the list for the sake of completion and usually its finished before I even notice and I get prompted to pick another research.

My approach is not really for speed runners, but it gets the job done... It's a strategy I've been using in all games that have static research trees... Probably because of how some games have huge penalties for leaping far ahead in the techtree in one direction while skipping all the other low level techs.



But from pure numbers... the Lab Research Speed is basically a useless research because it can simply be done with more labs/modules/beacons and therefore shouldn't even be in the tech tree or maybe work/scale somehow differently.

[mrvn]

How do those numbers look in marathon mode? Or with AAI, Bob's and Angels mods?

[BlakeMW]

How do those numbers look in marathon mode? Or with AAI, Bob's and Angels mods?

Generally in Marathon the numbers are much worse because research is so expensive compared with "stuff" (an example of stuff being extra beacons with speed modules) - stuff has a roughly 2-2.5x cost multiplier, research has a 8-10x cost multiplier (due to the packs being stuff, and requiring 4x as many of them). I don't think Bobs/Angels makes a big difference in the grand scheme of things - ultimately it comes down to making extra science packs for the lab speed research, vs making extra labs and extra productivity modules. The balance between science packs and labs+modules is not too different to vanilla - at least until you get the high end Bob's modules which are super OP and using them is always the optimal solution.

[Jap2.0]

For me, so long as I keep the (automated) research going on at most any speed, I have far more things researched than I'm doing. Other than the first few techs at the very start, and Advanced Oil Processing because you want it ASAP after getting oil running, I never, ever need something I've not researched hours before.

Same here... I'm almost always far ahead in research than what I'm currently doing in my factory (building stuff takes always more time than expected), except for Advanced Oil Processing which is always somewhat a stressful/frustrating experience which I'm not really looking forward to and which I'm happy when overcome.

That said, I always research the "Research speed" nodes, for no better reason than because they're there.

Me too.

Mostly I'm researching what's currently the cheapest to research (except if I need something desperately) so things take gradually longer and become gradually more complicated and when a new science pack gets added to the mix I usually expand the labs and entire infrastructure anyways to keep up... Eventually lab research speed becomes the cheapest research and that's when I knock it of the list for the sake of completion and usually its finished before I even notice and I get prompted to pick another research.

My approach is not really for speed runners, but it gets the job done... It's a strategy I've been using in all games that have static research trees... Probably because of how some games have huge penalties for leaping far ahead in the techtree in one direction while skipping all the other low level techs.



But from pure numbers... the Lab Research Speed is basically a useless research because it can simply be done with more labs/modules/beacons and therefore shouldn't even be in the tech tree or maybe work/scale somehow differently.

Yes, completely agree on this, I was speaking from a complete efficiency stqandpoint.

[Zavian]

Power requirements is also something that shouldn't be overlooked. When calculating the cost of additional labs, you also need to factor in the cost of providing power for those labs. But even if a particular research isn't worth it from a raw resource cost point of view, doesn't mean that research should be removed from the game.

There is even a use case when researching it makes sense. If you are building the largest longterm base you can, a base that is going to be ups limited, then less labs (and a more compact setup of labs in general) means less inserters and shorter average bot or belt travel distances, both of which mean your labs will end up being less of a drain on your ups.

[MeduSalem]

Power requirements is also something that shouldn't be overlooked. When calculating the cost of additional labs, you also need to factor in the cost of providing power for those labs. But even if a particular research isn't worth it from a raw resource cost point of view, doesn't mean that research should be removed from the game.

There is even a use case when researching it makes sense. If you are building the largest longterm base you can, a base that is going to be ups limited, then less labs (and a more compact setup of labs in general) means less inserters and shorter average bot or belt travel distances, both of which mean your labs will end up being less of a drain on your ups.

Honestly, in the grand picture I don't really think that power consumption for a few labs make a lot of difference. Neither the required infrastructure of a couple inserters/bots.


I never built a huge mega base with multiple rockets per minute before... but currently I'm planning on a 1000 science pack/minute factory (without expensive mode)... and I need about 1200 furnaces with PM3 und SM3 Beacons just to deliver enough Iron/Copper/Steel Plates. So if I need 50 labs or 25 labs doesn't much of a difference in the grand scheme anymore. I'd rather wish it would require less of the other blueprints which I've to copy-paste over and over.

[Zavian]

But at that point it is only going to cost you 30 seconds to actually do the last lab speed research, so why not ?

[MeduSalem]

But at that point it is only going to cost you 30 seconds to actually do the last lab speed research, so why not ?

Like I wrote in my first post... I'm researching it anyways just to knock it off of the list and to get it out of my face.

[Frightning]

I believe I did the math on this months ago when 0.15 was still fairly new and experimental. Basically only the red+green ones ever make sense, and even then, only if you are somewhere in the neighborhood of 80+ Labs going at once. (!)
They do however, beat moduling out Labs pretty fast, even for tier 1 modules (only takes like 20ish Labs iirc). Also, I am not certain of this, but I think that the Lab speed research is additive with speed modules, meaning it has fantastic synergy with Prod modules (countering their speed penalty additively, but I haven't tested this to confirm if it actually works this way). The blue science level lab research isn't worth it until you are consider at least tier 2 modules in labs, and only after lab count is getting into the dozens iirc. All the lab researches are generally better than tier 3 modules for increasing research, and in particular, are worth doing if you are considering using prod module 3s in Labs (not the ideal place for them, but if you've already moduled basically everything else in the chain, it's worth it to module the labs too for one more instance of prod multiplier stretching your materials).

[DaveMcW]

Labs (not the ideal place for them, but if you've already moduled basically everything else in the chain, it's worth it to module the labs too for one more instance of prod multiplier stretching your materials).

Labs are the second-best place in the game to put productivity modules, only behind the rocket silo.

[Cribbit]

Labs (not the ideal place for them, but if you've already moduled basically everything else in the chain, it's worth it to module the labs too for one more instance of prod multiplier stretching your materials).

Labs are the second-best place in the game to put productivity modules, only behind the rocket silo.

Only when using advanced sciences though.

[Koub]

viewtopic.php?f=16&t=23281