Lab Efficiency Research

[Tubig]

From what I can tell it is always better to build more labs than to research Lab Efficiency. Have I missed something?

Researching Lab Efficiency 1 would require 100 Red+Green Science packs. Total raw resources would be 750 Iron Plate + 250 Copper Plate. With those same resources I could build 16 labs. Lab Efficiency 2 costs more than 25 labs; Lab Efficiency 3 more than 268 labs; and Lab Efficiency 4 more than 387 labs.

Therefore, if I have less than 80 labs I get a faster research times by using those resources to build more labs than by researching Lab Efficiency. If my factory is large enough to be fuelling 80 labs, then most research tasks (94 out of 182) could be completed in less than a minute, many (136 out of 182) can be completed within 2 minutes. Only Logistic Robot Cargo Size 3, Logistic Robot Speed 5, Rocket Silo and Follower Robot Count 12+ would take more than 5 minutes. The longest research task (Follower Robot Count 20) would take little more than 16 minutes. All of the research tasks could be completed in less than 6 hours. At that point, it is doubtful that a further boost in research speed is needed.

A suggested fix would be to change Lab Efficiency so that it reduces the cost multiplier instead of the processing time. This would reduce the resources (science packs) needed for other research, making Lab Efficiency worthwhile.

[Khaylain]

I'm in agreement with this, seeing as efficiency would mean you don't spend as much on something. An efficient engine gives you more power for the same fuel usage as a less efficient one.
It would make more sense to have Lab Efficiency decrease the number of science packs needed by a certain % for each Lab Efficiency upgrade. And as Tubig has stated there is more incentive to build more labs rather than research the current Lab Efficiency, seeing as you don't need to spend the time and resources to make the science packs required for the current Lab Efficiency when you can just make more labs to process the packs faster.

[Koub]

This is totally correct. I thought I had already seen a discussion on that topic, but couldn't get to find it.
I also think this technology should be changed to "labs use less science packs"

[bobucles]

There already is a slight issue with the laboratory section of your base being so pitifully small. IMO lab research isn't needed in Factorio at all. If the player needs more research they should just like build more labs.

[seronis]

My assumption on this was that lab efficiency is something you work on after clearing all other low techs and before you have production of tier3/4 sci packs. Its so your labs arent sitting idle if you are slow to getting sci3/sci4 produced

[BlakeMW]

The maths which demonstrates lab efficiency is a waste of time is solid. Nevertheless a resource often unaccounted for is is player time/effort/attention. At some point you'll probably find it easier to just research the first two efficiency upgrades rather than expanding your labs. I wouldn't bother with the 3rd or 4th upgrades though - they cost an awful lot and also of the more expensive blue science packs.
I normally like to get about 32-40 labs and then research the first two efficiency upgrades in the mid-late game.

[Koub]

The fact is I see no use researching so fast : I alreadu never am able to use the outcomes of my research as I research them So I willfully limit myself to like 2 labs until my blue science production is flowing and automated.

[bobucles]

I alreadu never am able to use the outcomes of my research as I research them

That's the main reason why research speed is one of the more pointless techs. With the streaming research system it doesn't serve any real purpose. It's cheaper to build more labs and more fun to have a bigger lab system anyway.

[albatrosv13]

deleted

[BlakeMW]

The fact is I see no use researching so fast : I alreadu never am able to use the outcomes of my research as I research them So I willfully limit myself to like 2 labs until my blue science production is flowing and automated.

Maybe that's a play style difference? I put a high priority on bullet shooting speed, bullet damage, turret damage and robot follower count to get crazy powerful turrets and defender capsules early ish in the game (Like I definitely want gun turrets shooting bullets at 10 damage before small spitters evolve). Those techs ain't cheap in terms of research time, you can knock out all the oil processing techs in about the same time as Bullet Damage 3 which is a simple red/green tech:

Bullet Damage 3: 200 x 60s = 12000 lab seconds

Oil line:
OilP: 50 x 30s
Fluid Handling: 75 x 30s
Sulfur: 100 x 30s
Battery: 50 x 25s
Plastics: 100 x 30s
AdvancedP: 75 x 30s

Total: 13250 lab seconds

I can't imagine researching my war machine without at least 20 labs, 20 often feels too slow, 32-40 is better, and the efficiency upgrades are a nice addition too.

There are also other expensive techs, other military upgrades like shotgun and combat robot damage, and the whole logistic system. A player who largely abstains from the logistic system saves a lot of research, blue belts are so cheap in comparison with all the logistic upgrades. I happen to want both the unholy terror defender capsules and twinked out construction robots to repair my Tank - logistic robot speed is another tech which ain't cheap in terms of time. So I'm pretty much never short of really time expensive techs I want to research.

[bobucles]

But the energy saved while researching faster?

Labs consume 60kW. The energy savings of a faster lab competes against the cost of building 2 extra steam engines. It's a tiny benefit against an energy drain that is already very low.

Labs are a tiny portion of any player's base. Lab research means that small section gets to stay small. Small factories are sad.

[Xuhybrid]

It should become Lab Productivity. That way it's worth the cost investment because it saves you resources instead of time. Speed increases are easily handled by adding more labs.

[Tev]

It kind of degrades this entire section that thread like this can have 11 replies without even mentioning the correct math or what it is even based on.

Like, you know, modules.

Productivity modules 3 for 10 additional labs are not spare change (especially considering cost of the effectivity research). You can beacon them, but then cost of speed modules, beacons needs to be considered. Energy consumption also stops being rounding error. It still could win over cost-wise over highest tier of efficiency research, but it does not seem that clear to me . . . people should think more before posting in this section.

edit: also, Marathon / Deathworld settings. Those were already announced when OP was made, no?

[Koub]

You both answer to a 1 year old dead topic Though it's still revelant, a lot of changes have happened since.

[Tev]

Oops I missed the year of the post. My bad But in 0.15 it just doesn't seem so bad, and first 2-3 tiers were ok even before.

[BlakeMW]

It kind of degrades this entire section that thread like this can have 11 replies without even mentioning the correct math or what it is even based on.

Like, you know, modules.

Productivity modules 3 for 10 additional labs are not spare change (especially considering cost of the effectivity research). You can beacon them, but then cost of speed modules, beacons needs to be considered. Energy consumption also stops being rounding error. It still could win over cost-wise over highest tier of efficiency research, but it does not seem that clear to me . . . people should think more before posting in this section.

I don't think anyone was really thinking of productivity in labs before the really expensive Productivity and High Tech science packs came out. After all, in 0.14 you could just go kill a few more spawners for the alien science packs.

When it comes to productivity in labs, what the efficiency research has to compete with is speed beacons. So some super quick math: lab eff 4 costs 500 RGB beakers, a total raw resource cost of 27750. Speed3 modules cost in the ballpark of 2800 raw resources. Thus you can get about 5 Speed3 beacons for the price of lab eff 4, those beacons will provide a 100-200% speed boost for 12 labs. Lab eff4 increases the bonus from 90% to 140%, which is a (2.4 / 1.9) = 26% speed increase. So the beacons are being about 10x more effective for 12 labs. Thus if you had 120 labs with prod3 modules, the research would be a better investment.

I suspect then that speed3 beacons are generally going to be a much more potent investment than the blue-beaker lab eff techs - altough on the scale of the cost of modules, those techs aren't that expensive - just not very potent. It should be noted though that you can also cheap out and use speed2 beacons to get a even cheaper (except in electricity) speed bonus. You'd probably double-row the labs with speed2 beacons to get a +240% speed bonus for a lot cheaper. So +240% speed becomes the baseline, then the question is, what's more effective, getting +26% speed from the research, or piling more speed on by investing in 10x speed3 modules? This is probably the fairest possible comparison and it would be a bit more favorable to the research, because the +26% is cumulative with the +240%, whereas the upgrade to speed3 is additive on top of the +240%.

For Marathon/Deathworld because the research is 4x as expensive you can forgetaboutit competing with speed beacons, it's basically not going to happen unless you really crank up the resource richness settings allowing you to build a truly humongous lab setup, as with default settings you'd be pushing it running even 6 labs with a +240% speed bonus.

[Tev]

I don't think anyone was really thinking of productivity in labs before the really expensive Productivity and High Tech science packs came out. After all, in 0.14 you could just go kill a few more spawners for the alien science packs.

Please tell me it's a joke. It's a joke right? Why would you ever use productivity modules then, if not in the single most productive place in the factory? I mean I spent ~200 hours without using modules in labs too, but I didn't even know what a main bus is, and how to use trains properly . . . you can't really claim nobody used productivity modules in labs before 0.15, and I'd bet majority of people playing it for more than 100 hours uses them there.

When it comes to productivity in labs, what the efficiency research has to compete with is speed beacons. So some super quick math: lab eff 4 costs 500 RGB beakers, a total raw resource cost of 27750. Speed3 modules cost in the ballpark of 2800 raw resources. Thus you can get about 5 Speed3 beacons for the price of lab eff 4, those beacons will provide a 100-200% speed boost for 12 labs. Lab eff4 increases the bonus from 90% to 140%, which is a (2.4 / 1.9) = 26% speed increase. So the beacons are being about 10x more effective for 12 labs. Thus if you had 120 labs with prod3 modules, the research would be a better investment.

I suspect then that speed3 beacons are generally going to be a much more potent investment than the blue-beaker lab eff techs - altough on the scale of the cost of modules, those techs aren't that expensive - just not very potent. It should be noted though that you can also cheap out and use speed2 beacons to get a even cheaper (except in electricity) speed bonus. You'd probably double-row the labs with speed2 beacons to get a +240% speed bonus for a lot cheaper. So +240% speed becomes the baseline, then the question is, what's more effective, getting +26% speed from the research, or piling more speed on by investing in 10x speed3 modules? This is probably the fairest possible comparison and it would be a bit more favorable to the research, because the +26% is cumulative with the +240%, whereas the upgrade to speed3 is additive on top of the +240%.

Good point! Tier 4 is bad then (and should get bigger boost or smaller price), but I still don't know about the breakeven point - tier 1 is obviously worth the price, 2 probably too? 3? I can't look at the prices in game now.

For Marathon/Deathworld because the research is 4x as expensive you can forgetaboutit competing with speed beacons, it's basically not going to happen unless you really crank up the resource richness settings allowing you to build a truly humongous lab setup, as with default settings you'd be pushing it running even 6 labs with a +240% speed bonus.

Actually I'm still talking about the research in general. You can't really compare the price / effect ratio of tier 1 to tier 4. And when tier 1 (and 2, 3?) is good on it's own, changes should be imo only to higher tiers.

Not to mention that on Deathworld map you will NOT research beacons before yellow science (and frankly I feel like I have researched it too soon, having it before nuclear power), adding the cost of the research to the equation.

[BlakeMW]

I don't think anyone was really thinking of productivity in labs before the really expensive Productivity and High Tech science packs came out. After all, in 0.14 you could just go kill a few more spawners for the alien science packs.

Please tell me it's a joke. It's a joke right? Why would you ever use productivity modules then, if not in the single most productive place in the factory? I mean I spent ~200 hours without using modules in labs too, but I didn't even know what a main bus is, and how to use trains properly . . . you can't really claim nobody used productivity modules in labs before 0.15, and I'd bet majority of people playing it for more than 100 hours uses them there.

It's about Return on Investment. Basically in 0.14 there just wasn't enough tech in the game for prod3 modules to really pay off in labs, short of doing something contrived like only using a single lab to research all the techs. To play optimally you have to use the prod3 modules in the most productive locations first, meaning: rocket silo, green/blue circuits etc (as per productivity module math), so even if you can make it cheaper to complete the tech tree by making productivity 3 modules, labs would not be the first place you'd put them, you'd probably find better places for at least the first 20 prod3 modules and probably more if you're not doing something contrived like using a single assembler to complete the game (which someone has done I believe...)

Not to mention that on Deathworld map you will NOT research beacons before yellow science (and frankly I feel like I have researched it too soon, having it before nuclear power), adding the cost of the research to the equation.

That is true, there's no researching techs which require high tech beakers before high tech beakers, after all. In Deathworld I tend to put prod1 into the labs and just scale it up to be 30% bigger, while reserving the higher tier modules for more deserving recipes. By using prod1 modules you've already got 40% of the benefit of prod3 and at a tiny fraction of the cost. Next is prod2 modules, 60% the benefit at 1/5th the cost - scaling up still looks a lot more attractive than speeding up. Only once you've got prod3 in the labs does speeding up start looking more attractive, to get the most out of those super-expensive prod3 modules. If you look at the cost of scaling up prod1 or even prod2 setups, you can do a lot of scaling up for the cost of an labeff 3+ tech.

And beacons are completely essential for maximizing the use of prod3 modules, it's absolutely a priority tech if you're investing in prod3 modules and it's also the cheapest high tech-tech - only 300 beakers compared with 1000 for labeff3 and 2000 for labeff4, wouldn't surprise me if beacon tech is cheaper. In Deathworld it makes sense to focus heavily on productivity everywhere because there is so much time and resource investment for it to pay off. This makes the modules and beacons better value than at standard costs, but it doesn't do much to make the lab eff techs more attractive, because that speed bonus only applies to labs (which are still one of the lower priority places to put prod3 modules) while beacons can benefit all machines.

[Xuhybrid]

You both answer to a 1 year old dead topic Though it's still revelant, a lot of changes have happened since.

The fact that it is still relevant is why i didn't notice the year it was posted. We have mining productivity research but nothing for labs. Both machines can use modules.

[Ranakastrasz]

Indeed. the addition of productivity research is a good enough reason to revive this topic.

I personally always thought that it would make more sense if lab speed was lab productivity instead. I pretty much always used productivity modules in every lab I could manage, of as high a level as I could create.