I've been playing Factorio for a while, but never needed to log into the boards before, just been lurking. However, I was hoping some additional eyes on my expectations and computations for my TINS (There is No Spoon) run would be advantageous. Short history: My science is taking far too long due to shooting myself in the foot occasionally with some biter assistance and I want to check that my math is right.
I'm playing the experimental solo (when writing this: 0.15.16) and I'm trying to do it with the new science pack requirements. I'm also trying to do it on default settings, instead of the usual large starting area and overloaded resource patches for the achievement attempt. I'm enjoying the challenge of it, so no worries there. The new blueprints between games is absolutely wonderful in helping with this. Just figured I'd jump ahead a bit to some of the likely posts suggesting I turn the difficulty down.
So, onto my calculations and then my questions...
TL;DR: I calculate, if you can keep 4 Red Science Pack Assembler 2s as your limiter for science (assuming infinite Research Labs to absorb the output and instant science pack transfer to keep the calculations from becoming a log derivative due to transfer times) you can complete Rocket Silo science in just over 3 hours. How I came to that conclusion is involved and where I'd like the good folks here to double check me. I know I'm not using the golden ratio of 5/6/8, and that's simply because of my experience with the current mid-game and the supply throughput required for the new packs to keep up at the 1/1.33s pack rate.
So, for the first assumption, assume you can feed these beasts at full production rates:
4x Assembler 2 Red Science (R)
5x Asm 2 Green (G)
7x Asm 2 Blue (B)
4x Asm 2 Military Black (K)
6x Asm 2 Production Purple (P)
6x Asm 2 High Tech Yellow (Y)
Yes, that's a lot of assumption for factory throughput, but bear with me. If you can feed those beasts (particularly the late game ones) at full throttle, you get the following outputs:
R: 4 packs / 6.667 seconds or 1/1.6675s
G: 5 packs / 8 seconds or 1/1.6s
B: 7 packs / 10.667 seconds or 1/1.52s
K: 8 packs / 12.5 seconds or 1/1.5625s
P: 12 packs / 18.66 seconds or 1/1.555s
Y: 12 packs / 18.66 seconds or 1/1.555s
So, looking at research speed, the limiter is # of red packs produced per second for the science tree needed to get to Rocket silo. This means starting a new research every 1.6675 seconds at best. If it is assumed the 'trailing' time of a research is simply time to load up the next science packs to quick start the following one, the computation can ignore all but the final 60 seconds of the Rocket Silo itself.
So, what does this mean in practice, as far as I can tell. I'll start with a simple example, because if I've got a flaw in this part my entire process is flawed. Using the above listed construction, Lab Research Speed 1 requires 30 seconds x 100 science packs to produce. The packs/second above against the time required for a single pack to process shows that 18 labs are required to consume packs at an optimum rate, and after that the time to process is ignored if we have 18 or more labs available to consume the packs.
100 packs * production rate of the limiting science pack (1.6675 seconds for red, as green is faster) means that we can achieve Lab Research Speed 1 in 166.75 seconds, or roughly 2.77 minutes.
If this remains true, then the following tree is the minimums, barring defenses and railroads, needed to achieve Rocket Science with reasonable inclusions (like splitter belts and storage tanks): (Syntax: (Type of Limiting pack|# of limiting packs needed) Name)
R only items: (320 packs * 1.6675s) = 533s / 8.8min
(R|10) Automation
(R|20) Logistics
(R|30) Electronics
(R|50) Steel
(R|80) Automation 2
(R|10) Military 1
(R|10) Turrets
(R|10) Optics
(R|100) Fluid Handling
R/G Items: 2,545 packs - 4,243s / ~70.7min
(R|100) Research Speed 1
(R|200) Research Speed 2 (without these # of labs gets quite large)
(R|100) Toolbelt
(R|100) Oil Processing
(R|20) Military 2
(R|150) Sulfur
(R|50) Flammables
(R|200) Plastics
(R|200) Advanced Electronics
(R|250) Solar Energy
(R|100) Modules
(R|50) Speed Module 1
(R|50) Productivity Module 1
(R|50) Efficiency Module 1 (It's either this or a massive steam engine bank)
(R|75) Advanced Materials
(R|250) Concrete
(R|100) Engine
(R|100) Electric Engine
(R|100) Electric Energy Distribution
(R|150) Battery
(R|150) Accumulator
R/G/B Items: 850 packs - 1,417s / ~23.6min
(R|150) Advanced Oil Processing (Not needed for Rocket, but cracking is invaluable)
(R|250) Advanced Materials 2
(R|300) Advanced Electronics 2
(R|75) Speed Module 2
(R|75) Productivity Module 2
R/G/K Items: 520 packs - 867s / ~14.5min
(R|100) Explosives
(R|120) Rocketry
(R|100) Rocket Speed 1
(R|100) Rocket Speed 2
R/G/B/K Items: 800 packs - 1,334s / ~22.25min
(R|200) Rocket Speed 3
(R|300) Rocket Speed 4
(R|300) Rocket Speed 5
R/G/B/P Items: 600 packs - 1000.5s / ~16.75 min
(R|300) Speed Module 3
(R|300) Productivity Module 3
R/G/B/K/P/Y Items: 1000 packs - 1667.5s / ~28min
(R|1000) Rocket Silo
(I'm not retyping all that but Military 3, as an example, has a K limiter on it. It's just not needed unless you need the tank)
Adding these times up: 8.8m + 70.7m + 23.6m + 14.5m + 22.25m + 16.75m + 28m
arrives at: 184.6 minutes, or just over 3 hours.
Significant boosts to defenses (bullet only) costs ~1 hour of research time, and you can get the railroad up with a bit of boost to stack inserters and circuit networking in around 20 minutes.
So my main questions are the following:
Can anyone poke holes in my logic or my math?
Did I miss anything significant in the research tree?
If I get the chance (and there's interest), I'll post the materials throughput you'd need to keep that at full speed, I'm still digging through the build chains to break it down to (IP/CP/Oil/Coal)/second ratios.