Implementing Improved Robotic Controls

[Zeraan]

But would you be able to refit rather than producing more idle factories when you're below your max pop?

Hmm, I guess the logic for checking the "MAX" could be changed so that it checks if refit are possible, do that instead, otherwise build more factories. It'd be the most optimal use of industry.

[Zygot]

That is a good analysis Zygot. I think you leave out two important things, though: The return on investment of technology and colony ships. Let's ignore colony ships and say your only research project is Improved Ecological Restoration. When do you start researching it, and how many points do you devote to it? And with the opposite example where you can only build colony ships and factories, when do you start building a colony ship? Do you max out factories first? Lastly, there is the problem of tying up your population when overbuilding factories early on, preventing you from shipping colonists to new worlds without losing tons of production.

Good questions. I'm sure there's some math one could construct to determine the optimal answer, but here's my take.

First, I like your example of Improved Eco. It's a very good early waste reducer. I do like to open the fields very early, just for informational purposes (18 total RP split evenly over the six fields will do so), so let's say I see that. When do I start researching it? Well, it will be a priority for sure. Not only will it reduce my waste expenditures, providing a significant economic boost, but if I'm still in factory-building rush mode, it will increase the growth curve and get me to max even faster.

So what is its benefit anyway? Well, it reduces the waste penalty of a working factory from 1/2 BC to 1/3 BC (subtracted from the 1 BC gross production). That improves net factory production from .50 to .66 BC/factory, or a relative 33% improvement. Not bad. So the production of every 200 factories goes from 100 BC to 133 BC a turn.

In addition the workers produce more inherently themselves, from .53 BC/worker to .65 BC/worker, or a relative 23% improvement. Again, not bad. So the production of 100 workers goes from 53 BC to 65 BC.

In sum, Improved Eco gives a theoretical production boost to a maxed 100 pop/200 factory home world from the baseline 153 BC to 198 BC, or 29% better. Hell yes, I'll take one of those, please!

Now unfortunately it will typically cost 1000 BC to research Improved Eco. It's hard to say because you get the continuing research investment over time, but to balance that you have to put more than the base cost to get a reasonable chance of discovery on any given turn. So let's just call it 1000 BC.

On a superficial analysis, it will cost you 1000 BC to grant the benefit of a 45 BC/turn improvement on your fully maxed home world. Now think about various scenarios. If you are not maxed on your home world, let's say you're half that at 50 pop and 100 factories, it would cost you 1000 BC for a 22.5 BC/turn improvement. See where this is going? The bottom line is that:

Improved Eco Restoration is significantly more valuable, and gets much cheaper on a relative basis, as you have more factories and more production.

At game start the cost/benefit ratio is not worth it, but if you have several maxed planets you're essentially flushing free money away not to get it.

Therefore, it doesn't make sense to start researching it too early; you don't want to sink a large fixed cost into a small return, you want to sink that same fixed cost into a large return!

So in my games, I will still factory rush; I think climbing the "hockey puck" growth curve is essential. To actually answer the question:

My typical approach is to, besides building a few Scout Mk IIs, to build factories non-stop until I naturally reach the point at which they cannot be worked, at the 2x population mark, typically a situation like 60 pop/120 factories or 75 pop/150 factories. At that point, it's time to start putting that production to use. Maybe start in on the Improved Eco; make it my first tech, or one of the first ones, certainly above something like Hand Lasers or Force Field II! I will definitely aim to have it by the time I have my third colony chugging along.

But if I have a good opportunity with another tech, like Range 4 or Range 5 to settle something juicy or open up some juicy planets, then I will prioritize those instead. Those will give even better returns than waste reduction.

Now for your other question, about when to build colony ships, I use the same logic. Factory rush until the pop/robotics equivalancy point, then build a colony ship.

In any given game, in the early years, priority is always given to colony ships over research.

I do not like overbuilding factories. Doing so is a waste of money. In the early game, it's OK to do some micromanaging to ensure all built factories can actually be worked. Later in the game, I don't care as much, as long as it's not egregious.

[Catwalk]

If there was no research bonus, it might be possible to calculate the best approach. With the research bonus, it becomes a lot more complicated. Think about it this way: If you want IER at some point, you can get 3x the research out of your BC by riding the curve all the way until completion. I'm going to try to sim this out for you in a spreadsheet, assuming 100 starting pop to keep things simple. Maybe later I can add in pop growth to get a more practical application. I think that early IER research is critical to getting a good start, and I think I can prove that it's worthwhile investing in it much sooner than you're suggesting. On the assumption that you're fully utilizing the research bonus, of course.

Also, if you use fewer paragraphs your long posts are easier to read

[Catwalk]

Okay, made a very crude simulation under the following two scenarios (both assuming 100 starting pop):

Case 1
I build up to 200 factories then max research
200 factories reached on t19
Total accumulated production output on t27 = 3179
Total accumulated bonus on t27 = 602
Total RP on t27 = 2081


Case 2
I put 63 BC into research on t1 then put 7.5% of accumulated RP into research on subsequent years
200 factories reached on t27
Total accumulated production output on t27 = 2932
Total accumulated bonus on t27 = 1154
Total RP on t27 = 2372

No rounding takes place in either scenario, so this is not entirely accurate. I feel that the difference is significant, though. If you factor in getting IER one turn sooner in case 2 (both have 2080 RP on t26), you get an extra 33 BC in that scenario. It is likely case 2 gets the tech sooner, as it gets into the breakthrough percentages 2 turns earlier.

For any case where I delay the initial IER investment or don't ride the bonus consistently, I get worse results. Would you always get IER before your first colony ship, or will you sometimes get the colony ship first?

[Zeraan]

I got a suspicion on how things work internally, based on how missile bases work. What happens when the following scenario occurs:

A planet has 100 pop and 200 factories, and you just recently discovered Improved Robotic Controls III. You don't assign building new factories, but rather, you send 50 people out from the planet.

Does the game know that the planet is at Robotic Controls II (only 100 factories are used), or does it know only that it needs to pay 500 BC before it can build more factories (10 + 5 for III control), and treats all existing factories as III (150 factories are used, despite it not being refitted yet)?

[Zeraan]

I got a suspicion on how things work internally, based on how missile bases work. What happens when the following scenario occurs:

A planet has 100 pop and 200 factories, and you just recently discovered Improved Robotic Controls III. You don't assign building new factories, but rather, you send 50 people out from the planet.

Does the game know that the planet is at Robotic Controls II (only 100 factories are used), or does it know only that it needs to pay 500 BC before it can build more factories (10 + 5 for III control), and treats all existing factories as III (150 factories are used, despite it not being refitted yet)?

Bingo! I conducted tests myself. Here's the results:

100 Pop, 200 Factories on home world, no ships or bases, two planets in empire - Robotic Controls II:
253 (253) production (200 factories + 53 from population)

On turn of transferring 50 population:
203 (253) production (50 BC are deducted for transfer)

Turn after transferring 50 population, with 52 pop (+2 growth):
131 (131) production (100 factories + 31 from population? Shouldn't it be 21 or 22ish? Where's the extra 10 from?)

On turn of discovering IRCIII, and declining investing into new factories
253 (253) production (200 factories + 53 from population)

Same said turn after ordering 50 to be sent
203 (253) production (50 BC are deducted for transfer)

Turn after transferring 50 population, with 53 pop (+3 growth):
187 (187) production (150 factories + 27 from population?)

As you can see, the game sees that there's 200 factories, and don't store whether or not they're refitted. Apparently it only keeps track of how much is invested to simplify things on code end. My guess was correct. Wonder if this affects any strategies?

[WhiteMage]

Your math is wrong. 52 pop uses 104 factories out of 200 at IRC II. 52 pop and 200 factory produces 52/2+52*2+e=26+104+e=130+e production, where e is production due to Planetology level. It is a percentage of total single planet's production. For 250 it is 3, for 130, it is 1. For 185, it is 2. With higher planetology, e gets bigger. It does not behave perfectly likely due to rounding errors.
There is similar math error in your other example.

And yes it does impact strategy, which is good.

[Zygot]

Catwalk, I've only used the optimal research method once in a game, which was one of my Imperium reports. It worked well, but I found the process to be an excessive amount of micromanagement.

You say the player can get 3x the research out of the invested BCs, is it really that much? Let's say a tech like Improved Eco is 1000 BC, if I use the optimal method, how much would I have actually spent, and how much would have been gotten "free"? I'm a little confused by your two scenarios posted above.

[Catwalk]

I recommend it for the first 2-3 techs. Because of compound interests and the value of those early techs (IER, IT+10, range 4 or 5) you gain huge rewards for that. Here is another example:

In case 1 you invest 40 BC on turn 0, then invest 7.5% of accumulated RP each following turn.
In case 2 you invest 45 BC each turn (minus 12 BC total to make numbers match).

In both cases you reach 2473 RP in 22 turns.
Case 1 costs you 851 BC
Case 2 costs you 978 BC

Furthermore, in case 2 you commit more resources earlier on. This gives case 1 an additional bonus by being able to devote more resources to factories at an earlier time.

[Zeraan]

Your math is wrong. 52 pop uses 104 factories out of 200 at IRC II. 52 pop and 200 factory produces 52/2+52*2+e=26+104+e=130+e production, where e is production due to Planetology level. It is a percentage of total single planet's production. For 250 it is 3, for 130, it is 1. For 185, it is 2. With higher planetology, e gets bigger. It does not behave perfectly likely due to rounding errors.
There is similar math error in your other example.

And yes it does impact strategy, which is good.

Oops! I forgot the extra 2 or 3 population that grew the past turn. Planetology level is at 1, I only devoted research into Computers for testing. From what I understood, you start at 0.5 production per population at Planetology level of 0, and caps out at 2 at level 99. So level 1 is like 0.515 production using this formula: ((1.5 / 99) * planetology level) + 0.5.

So in first wrong math, 52 population would produce about 26.78 plus 104 from factories, totalling 130.78, which if rounded up, is 131.

And in second wrong math, 53 population with IRC III would produce about 27.295, plus 159 from factories, totalling 186.295, which if rounded up, is 187 production.

Those numbers now add up

Anyway, we now know that when you research a new IRC, you're actually wasting some production if you leave people on a full planet, better to send them off to new planets to help boost their production. The question is, how much before it's too much? We have to consider that sending transports deducts from the planet's production as well. Ideally, we'd want the production to match pre-transport.

So if we have a planet with 100 population and 200 factories, we know that it has 253 production. We want to send out people to new planets, but keep the 253 production, after we research IRC III. The minimum population now required to keep the 200 factories operating is 67. So that's 200 production from the 200 factories. 67 population produces 34.5 production, which rounds up to 35 production. We now have 235 production, a slight drop from 253 (18 production loss), if we send out 33 population.

If we go from IRC III to IV, with the same 100 population planet. Assuming that we have 300 factories, we need 75 population to keep them running, so we can send out 25.

I can see that this is a diminishing return. The higher IRC you research, assuming that it's only one level above your current one, the less population you can send out. Biggest bonus is at start of game, from IRC II to III. Of course, going from II to IV or III to VI would be huge. Also, terraforming will impact the population count as well.

The formula for calculating how many people you can transport is P - (F/L) where P is current population, F is current factories, and L is highest IRC level you have.