View source for 322,059 resources by Madman, Apr 2009

Some details, and a brief 'how to'

== Universe and race ==

240 planets in a small, packed, max minerals, AccBBS, no random events. I'd consider it an average hab draw. The only pregenerating advantage I did allow myself was to create games until I had a universe where the starting planet was close to centered, as trying to fill from an edge or a corner is a huge expansion hit.

* CA, zero points left over,
* IFE, TT, ISB, NRSE, OBRM, NAS
* 0.58g to 1.72g
* -72C to 72C
* 30mR to 72mR
* 20% 1 in 14
* 1/1000
* 15/8/10 G-box
* 10/3/10
* Bio cheap, rest expensive, no start at 3

This is of course a broken race (look at the narrow hab, expensive tech and poor mines), only suitable for testbedding.

== Overall Strategy ==

The main thing is to keep the total population growth rate as high as possible, building space docks, colonisers and freighters just as they were needed, but not letting anything get in the way of pop growth. Also get bio up as fast as possible - the maximum growth of a planet is proportional to the square of the hab.

The other thing is that in a real game, the econ focus is the 'resource integral' - how many usable resource you get over time. In this, resource integral is important till the techs are done, and after that the focus is 2450 - it doesn't matter what the resources are at 2449, and that growth might slow growth way down at 2451.

Another important thing is to do things 'just in time' - with a fast resource growth rate, one resource now is worth several resources later, so don't spend resources on things that are not going to be used immediately (OK, this applies somewhat for real games too)

This does require intense micromanagement. I think the reason this attempt did so much better than the previous ones is that I micromanged all the way till 2450 - a lot of the other attempts seemed to not do any after 2438 or 2440, which would cost them a lot of resources.

== Scouting and expansion ==

It important to have pop (particularly breeder planets) spread relatively evenly over the whole universe as quickly as possible. Otherwise there is the problem of having central planets with big trips for all its expansion, and lots of greens near the edge with no breeders nearby. With the sort of effective growth rate I was going for, leaving it 5 years means transferring twice as much pop, so I'd go for that 80% planet 5 years travel away before going for that 70% that is only 2 years travel. The means scouting and expanding really aggressively.

My start:
* 2400:
** turn research down to 0%, set research to construction,
** autobuild lots of factories, mines,
** set the scout and change of heart scouting right away
* 2401:
** build as many scouts as you can in one turn,
* 2402:
** send off the scouts,
** set research to a percentage that allows construction 3 in two years,
* 2404:
** start building colonizers and medium freighters, and the expansion starts.
** build more scouts as required or resources available to cover any 'holes'. 

I built 15 scouts, and had just about all of the universe scouted by 2420.

== Planet hold points ==

For most of the test, I held planets to 25%. Only very late (no planets with less than about 40% hab) did I increase this briefly to 33%, then to 40% right to the end (I would have picked a higher number than 40 if I'd needed more factories). 40% was chosen because it has very little less growth than 33%, and allows more factory/mine building. I'd also boost a planet if it was below 25% full and had a breeder with nothing better to do nearby.

After I had all the planets to 25%-40%, any remaining pop got parked in freighters, and got dumped on planets as a waypoint 1 task in turn 50. This can be seen in how the growth slowed down after 2440:

* year	rsrc	pop	 av growth
* 2400	140	125	
* 2410	745	648.1    1.1789%
* 2420	3659	2717.9   1.1541%
* 2430	16000	10733.8  1.1472%
* 2440	60000	43105.9  1.1492%
* 2450	322059	129015.8 1.1159%

== Tech gain ==

First the obvious - get to con 3 in time to expand.

After that, it's a close call, but I went for bio 9 (total terraform 10) ahead of con 4 - the idea is to do the best with pop that I could.

I had a simple system for after that - ignore research completely (set to 0%) until I could get the next level of total terraform in 3 turns at current production with a bit to spare. Then I'd turn all the factory and minebuilding off for three turns (hence the wiggles in the resource history), but keep the spacedock/freighter/colony machine ticking over (that is important!), letting a planet slide above the hold point if I needed extra resources on the third turn, or alternately setting my weakest (lowest number of factories) colonies to autobuild factories/mines for the third turn if I had surplus resources.

Somewhere in the middle (between bio 13 and bio 17), I increased con to 8, allowing large freighters - there is a lot of pop moving that needs to be done and large freighters are slightly resource cheaper and hugely ironium cheaper than privateers. In a previous try at this I kept running low in ironium because of all the privateers I needed to build.

The results were something like:

* 2404: con 3 (medium freighters)
* 2414: bio 9 (TT10)
* 2415: con 4 (privateers, space docks)
* 2426: bio 13 (TT15)
* 2430: con 8 (large freighters)
* 2434: bio 17 (TT20)
* 2439: bio 22 (TT25)
* 2442: bio 25 (TT30)

After 2442, techs didn't matter, so I started focusing on what would get the highest resources at 2450, not worrying about the intervening years (things like building mines before factories where appropriate).

== Turn Order ==

My sequence each turn after the scouts were all sent out was something like:

* deal with any tech advances turning on autobuild factories/mines for all non-full planets if just got one,
* manually unload all the pop where a freighter arrived at the same time as a colonist,
* view the planets in order of capacity (adjusting the percentages for the hold point),
** for all the planets with 17+% capacity, amke sure they have a stardock building and at least 30 mines (for freighter construction)
** for planets with 21%-25% capacity, make sure a freighter is arriving next turn or build one,
** for planets over 25% capacity, ship off surplus pop (and germ if possible) to an under/unpopulated world, and make sure they have a freighter for next year
* view planets by reverse factory number, turn off autobuild to make sure that factories/mines aren't built past the hold point,
* view ships in reverse order of cargo, make sure they are all doing something, and have the quickdrop set,
* view ships in order of fuel, then ETA. Those that are not fully fuelled and are sitting still get sent to a breeder,
* view ships in reverse order of ETA, forward through the non ETA-1 ones and check they are getting there in minimum time,
* decide whether to go for another tech advance (is total resources greater than next useful tech divided by three?): if so, then go turn off autobuild factories/mines for planets.

Repeat this (with some variations) about 40 times to get a good score!

== End Game ==

I did two things at the end that you wouldn't do in a normal game.

Firstly, I had been building up a lot of pop in freighters rather than slowing down the growth rate (you wouldn't do this in a real game, as it hurts the resource intergral), that all got landed on planets that had surplus factories as a waypoint 1 task in 2450.

I also had some planets that weren't going to have complete factories built. In 2449 (in some cases 2448), I lifted any population that was going to be factoryless and shifted it (waypoint 1 task in 2450) to planets with surplus factories.

== How to get a better result ==

The biggest improvement would be by increasing the size of the universe to medium or large, and not having to colonize the poorer planets so quickly. It seems like cheating though to not use the 'standard' testbed.

Secondly, I didn't have nearly enough factories. I'd change the settings to one of:

* 15/8/11 G-box, 9/3/11
* 15/8/12 G-box, 8/3/12
* 15/8/13 no G-box, 10/3/10 (probably this one)

or get the points by removing one of IFE/NRSE or ISB, or adding CE:
* removing IFE/NRSE has the problem of really slowing down expansion and scouting right at the start, and takes till prop 8 or prop 9 (remember prop is expensive) to get a good warp 9 engine.
* removing ISB is a great points mine that doesn't hurt at the start, but might cause a freighter building crunch when the first wave of colonies become breeders. Also, the pain and micromanagement of refuelling points costing 600 resources to build rather than 80 is going to be there for the rest of the test.
* cheap engines might be another option for masochists. One thing it really messes up is that freighters can sometimes arrive at a planet before a colony ship.

Thirdly, with the experience of doing this once, I could probably make some better decisions during the test. There's probably not a huge gain there though.

Lastly, random events. Turning random events on and getting a lucky mystery trader (getting to a bio breakpoint early, or the alien miner near the start), wormhole (unlikely with only the homeworld having a scanner), or just getting the right techs from early colonies. I don't like this as it makes the result irreproducible.