Home World Forum
Stars! AutoHost web forums

Jump to Stars! AutoHost


 
 
Home » Primary Racial Traits » AR » Energy at normal / expensive
Energy at normal / expensive Wed, 28 May 2014 14:09 Go to next message
mrvan is currently offline mrvan

 
Officer Cadet 1st Year

Messages: 220
Registered: May 2014
I was re-reading some of the other posts and there was a discussion going on about whether energy should always be taken cheap.

Intuitively, the increasing tech costs (which means that 50% cost means a lot less than 2x the tech level, more like 1 or 2 levels higher) combined with the square root effect makes energy cheap not terribly attractive. In most games, I research energy up to tech 4/10/16 at various moments, but hardly ever much higher. At energy normal this costs around 107090 (cumulative, not taking penalties for other tech into account), which at cheap would give level 19. This would give me sqrt(19/16)=1.089 or 9% less resources.

To get some numbers, I've put together a spreadsheet that calculates cumulative tech cost and compares where normal/expensive/expensive@3 stands compared to cheap for the same investment in energy tech. The graphs at the bottom give tech level equivalence and most importantly the resource hit per cheap tech level. Until around (cheap) tech level 10, growth is hit for normal and expensive, with a penalty of around 25% for normal and 40-50% for expensive. Normal settles at a ~10% resource hit, expensive around 20%. Taking expensive plus start at three ameliorates the slow start, starting more quickly than cheap for the first couple levels and never having much more than a 20% hit.

Spreadsheet: https://docs.google.com/spreadsheet/ccc?key=0Atz2h7DLvPeHdHV BZGY3VThqNktzZ2J3MjI2WTd1bVE&usp=sharing
Resource hit Chart: http://i.imgur.com/yzvssDS.png

In the RW, taking energy normal instead of cheap yield 84 points, taking expensive@3 also yields 84 points, and expensive alone yields 144 points.

I'm assuming here a fairly default race, IFE+ISB+NAS, 1i 2 narrow 1 in 6, 15% growth, 1:1000, C+W cheap, P+El+B expensive.

84 points can buy you around 6 hab clicks (1:6 -> bad 1:4) or 1% growth and 1.5 hab clicks (1:5)
144 points can buy you 9 hab clicks (good 1:4) or 1% growth and 4 hab clicks (1:5) or 2% growth and 1? hab click.
(144 plus 3 hab clicks (1
...



[Updated on: Wed, 28 May 2014 16:14]

Report message to a moderator

Re: Energy at normal / expensive Wed, 28 May 2014 16:24 Go to previous messageGo to next message
skoormit is currently offline skoormit

 
Lieutenant

Messages: 665
Registered: July 2008
Location: Alabama
Very interesting stuff there.

You are missing one piece of the analysis: the compounding effect of the "resource hit" on the ability to spend equivalent resources in tech.

For example:
With energy cheap, lets say you spend 105 resources in 2400 to get to En3.
With energy normal, you need 210 resources to get to En3. But you don't have that many in year one. Let's say you spend the same 105 as energy cheap did, and you get to en2 and make some progress toward en3.

In 2401, energy cheap has more resources to spend than energy normal.
To get to en3, energy normal has to spend another 105 resources, but that represents a bigger portion of total resources than it would for energy cheap. A 22% bigger portion, in fact. It would be like energy cheap spending 128 resources.

You have to factor in that compounding in order to complete the analysis.

Another way to look at it: you've examined total resource cost implications, but you haven't examined time implications. It may be true that an energy normal race that spends 80040 resources on Energy gets to level 15, while the energy cheap race that spends the same amount gets to level 17. But it's not true that the energy normal race can spend that many resources as soon as the energy cheap race can. When the energy normal race is finishing level 15, the energy cheap race will be farther ahead of it than your current analysis indicates.

You also aren't factoring in the increase in cost due to other techs being researched (5/10/17.5 for every En level for every tech level researched). That's a small effect in the overall analysis, though.



What we need's a few good taters.

Report message to a moderator

Re: Energy at normal / expensive Wed, 28 May 2014 16:36 Go to previous messageGo to next message
mrvan is currently offline mrvan

 
Officer Cadet 1st Year

Messages: 220
Registered: May 2014
Thanks!

You are right about the compounding effect, but whatever you buy for the RW points will also contribute to growth somehow, so that is quite difficult to determine accurately. An extra habitable planet within 1 jump or an extra growth% also increases resources. So I'm not sure what do to except for testbedding for this, unless you have a good idea for calculating the growth effects of extra hab/%?

(intuitively, I had hoped to compensate for expensive energy with 1:900, but my god that is expensive...)

I added the extra 8 tech levels for exp@3 on afterthought and it had no effect whatsoever, so the effect is quite minor indeed. Given that most players will research energy to 5 or so before researching anything else and even then only research say con 4 weap 5 before moving to energy 10, so I don't think the extra 90/2=45 resources will make that much of a difference.

If I have time I might try to testbed an 1:x (1000<x<2500) efficiency 13 or 14% 2i. It is quite nice how quickly every planet becomes 100% with 2i+narrow, but tricky to get to work...

Report message to a moderator

Re: Energy at normal / expensive Wed, 28 May 2014 16:56 Go to previous messageGo to next message
skoormit is currently offline skoormit

 
Lieutenant

Messages: 665
Registered: July 2008
Location: Alabama
Testbedding can be useful, but for comparisons like these there's just too much noise from the randomness (like you said, 1 more green within 1 hop makes a huge difference).

I've always had a hard time coming up with a useful way to calculate the value of hab settings and econ settings on a single scale.

A monte carlo sim might work:
Generate random hab values for a huge number of planets. (someone knows the random algorithm, right?)
Then iterate the following:
1) select from the set a random collection of X, Y, and Z planets as 1hop, 2hop, and 3hops from the HW. (universe size/density + geometry should tell you the expected values of x, y, and z, modulo corners--so assume universe remapped to a circle and hw at least 245ly from edge of circle).
2) pass those 3 planet sets and a set of hab/production settings to an algorithm that determines the optimal expansion pattern among these planets for this race and spits out a "resources by year" theoretical result (smells crunchy, but probably doable, especially if you handwave some stuff that doesn't move the needle much)
3) dump those theoretical results into a database of some sort


After you've got enough data, compare the results for different hab/production settings.

I'd really love to know how long it takes Total Terraforming to pay off....




What we need's a few good taters.

Report message to a moderator

Re: Energy at normal / expensive Thu, 29 May 2014 19:41 Go to previous messageGo to next message
Altruist is currently offline Altruist

 
Commander

Messages: 1068
Registered: August 2005
Location: Berlin
mrvan wrote on Wed, 28 May 2014 20:09
I was re-reading some of the other posts and there was a discussion going on about whether energy should always be taken cheap.

Very interesting approach and question.
skoormit
You are missing one piece of the analysis: the compounding effect of the "resource hit" on the ability to spend equivalent resources in tech.

Certainly true.

Here another perspective, the reality check:
Playing AR is difficult in a real game. Once you have death stars and a huge amount of remote miners your AR is booming and a strong contender for the win. Mostly you won't make it that far because you get crushed by other players before.
So to make an AR competitive, you have 3 choices:

a) AR on its own: needs to be faster and stronger. But so far nobody was really able to cope with the problems of
# needing 3 main techs to research: en-con-weap
# producing all the remote miners you need while the other players build those fleets to kill off your starbases

b) Play in an environment better suited for ARs:
# ARs only
# lots of space and time to develop and get ready

c) Play an AR with a non-AR-ally plus some space and time to develop:
This is the situation AR is played most often. An AR as an ally in a long game is a very good ally and usually brings in CHEAP energy and minerals in the endgame. Due to tech exchange between allies it is quite mandatory to take energy cheap for an AR.
Quote:

[...]
84 points can buy you around 6 hab clicks (1:6 -> bad 1:4) or 1% growth and 1.5 hab clicks (1:5)
144 points can buy you 9 hab clicks (good 1:4) or 1% growth and 4 hab clicks (1:5) or 2% growth and 1? hab click.

It would be interesting to see 2 ARs competing: one with en cheap, the other with normal and the pts spent elsewhere. The one with energy cheap should have a much faster start but would be forced to do something with his headstart... early attacking? Otherwise his advantage will get smaller and smaller during midgame...
...



[Updated on: Thu, 29 May 2014 19:43]

Report message to a moderator

Re: Energy at normal / expensive Wed, 04 June 2014 10:33 Go to previous messageGo to next message
mrvan is currently offline mrvan

 
Officer Cadet 1st Year

Messages: 220
Registered: May 2014
Although testbedding an en-expensive race gave disappointing results, I realized another option: en expensive instead of cheap almost completely pays for TT, which reduces terraforming costs by 30% and will in the very long run offer additional terraforming possibilities. Maybe with the extra 60 points for start@3 this might not be too sluggish at the start, as that not only gives a small boost to starting production but also obviates the need to spend early turns on expensive prop and bio to get them to 2 (for mizer + terraform +/- 7)

Report message to a moderator

Re: Energy at normal / expensive Wed, 04 June 2014 10:39 Go to previous messageGo to next message
mrvan is currently offline mrvan

 
Officer Cadet 1st Year

Messages: 220
Registered: May 2014
(sorry dupe)

[Updated on: Wed, 04 June 2014 10:39]

Report message to a moderator

Re: Energy at normal / expensive Wed, 04 June 2014 15:29 Go to previous messageGo to next message
skoormit is currently offline skoormit

 
Lieutenant

Messages: 665
Registered: July 2008
Location: Alabama
mrvan wrote on Wed, 04 June 2014 09:33
Although testbedding an en-expensive race gave disappointing results, I realized another option: en expensive instead of cheap almost completely pays for TT, which reduces terraforming costs by 30% and will in the very long run offer additional terraforming possibilities. Maybe with the extra 60 points for start@3 this might not be too sluggish at the start, as that not only gives a small boost to starting production but also obviates the need to spend early turns on expensive prop and bio to get them to 2 (for mizer + terraform +/- 7)


Be very careful about TT in a game against humans.

TT can seem like a very good investment in the long run. But in multiplayer games you will be fighting for planets. If you have less available econ because you bought TT, you will win fewer planets. If you win fewer planets, you are less likely to survive long enough for TT to pay off.



What we need's a few good taters.

Report message to a moderator

Re: Energy at normal / expensive Thu, 05 June 2014 12:52 Go to previous messageGo to next message
mrvan is currently offline mrvan

 
Officer Cadet 1st Year

Messages: 220
Registered: May 2014
skoormit wrote on Wed, 04 June 2014 15:29
mrvan wrote on Wed, 04 June 2014 09:33
Although testbedding an en-expensive race gave disappointing results, I realized another option: en expensive instead of cheap almost completely pays for TT, which reduces terraforming costs by 30% and will in the very long run offer additional terraforming possibilities. Maybe with the extra 60 points for start@3 this might not be too sluggish at the start, as that not only gives a small boost to starting production but also obviates the need to spend early turns on expensive prop and bio to get them to 2 (for mizer + terraform +/- 7)


Be very careful about TT in a game against humans.

TT can seem like a very good investment in the long run. But in multiplayer games you will be fighting for planets. If you have less available econ because you bought TT, you will win fewer planets. If you win fewer planets, you are less likely to survive long enough for TT to pay off.


Agreed (although I have limited pvp experience, and what I have is a decade old). However, for AR terraforming IS economy. In a way, terraforming are factories, ie you buy more production capability. With TT terraforming is cheaper, so you get planets contributing to research/shipbuilding earlier or at more capacity. Then again, a standard 1i AR that needs to do full terraforming (2x15 clicks) pays 3000 resources, which is a lot for a red planet, but a lot less than even a HG would pay for full factories (10-15k depending on settings).

Assuming 250K pop, energy 16, and 2% hab gain per terraforming, 100 resources of terra yield 12.6 extra resources, so equivalent to (default) factories costing around 8. At 250K pop and energy 10, this is only 10, so factories cost 10. TT changes this to factories cost 5.6 ~ 7, which is a pretty good deal. Problem is of course that you don't get them at the HW, and only limited at other worlds, say 10 clicks on most of the worlds you colonize early, which equates around 100-150 factories max. depending on tech level and population. The difficulty with AR is that everything depends on everything else because of the sqrt non-linearity.

What about placing energy outside the sqrt but reducing the capacity of US and DS to 1.5M and 2M?

Report message to a moderator

Re: Energy at normal / expensive Fri, 06 June 2014 00:56 Go to previous messageGo to next message
skoormit is currently offline skoormit

 
Lieutenant

Messages: 665
Registered: July 2008
Location: Alabama
Terraforming is economy for everyone, because ultimately econ is capped by capacity.

TT is very, very expensive. I'm not saying it can't work, I'm just reminding you of the hidden cost: planets you don't gain control of because you didn't have enough early power to win the race. Every such planet you don't win is a planet that a neighbor gains. That's a difference of two planets in relative power. How many planets do you have to lose out on before TT is no longer a net win even in the long run?



What we need's a few good taters.

Report message to a moderator

Re: Energy at normal / expensive Fri, 06 June 2014 12:29 Go to previous messageGo to next message
XAPBob is currently offline XAPBob

 
Lt. Commander

Messages: 957
Registered: August 2012
but it's direct economy, even without growing more colonists for AR - of course they don't get the max pop boost, but they get the economy now, and more growth. others get the growth, but they have to wait for the economic return.

Report message to a moderator

Re: Energy at normal / expensive Fri, 06 June 2014 14:10 Go to previous messageGo to next message
skoormit is currently offline skoormit

 
Lieutenant

Messages: 665
Registered: July 2008
Location: Alabama
XAPBob wrote on Fri, 06 June 2014 11:29
but it's direct economy, even without growing more colonists for AR - of course they don't get the max pop boost, but they get the economy now, and more growth. others get the growth, but they have to wait for the economic return.


By the mid-game, terraforming is tantamount to direct economy for most +f races, assuming competent pop and mineral management. When breeders start sending pop to the planets that have greenlined (or greylined) factories, that new pop operates factories the year it arrives (or the year after if you don't build the factories ahead of time). Non-breeders start terraforming when they fill up, so they can hold more people and operate more factories. If you are terraforming non-breeders before they are at 100% population with full factories and mines, you probably don't have enough breeders (or got rather unlucky with your germanium draw).

I'll agree that in the early game an AR doesn't have to wait for terraforming to pay off. But if you want early game power TT is about the worst way to spend that many RW points.



What we need's a few good taters.

Report message to a moderator

Re: Energy at normal / expensive Sat, 07 June 2014 08:51 Go to previous messageGo to next message
XAPBob is currently offline XAPBob

 
Lt. Commander

Messages: 957
Registered: August 2012
You still have to grow the pop/build the factories as well as TF. The TF directly affects the AR hab which is not only in the resource forumla, it's outside the sqrt.

with planning you can schedule those factories to be ready in time for more people to arrive, but you're still buying factories as well as TF.

Report message to a moderator

Re: Energy at normal / expensive Sat, 07 June 2014 14:20 Go to previous messageGo to next message
mrvan is currently offline mrvan

 
Officer Cadet 1st Year

Messages: 220
Registered: May 2014
There are 5 variables that (directly) determine AR economy:

- #planets
- hab
- population
- energy tech
- divisor

Each of them are more of less directly influenced by the RW settings. Initial growth is mainly affected by pop, divisor, and energy tech; ramp-up on new planets is also determined by hab, and #planets is extremely important as it is the only variable outside the sqrt that is in principle unbounded. Terraforming is great but you can't terraform the HW and you can only terraform a limited number of clicks on other words.

What makes optimizing AR so difficult is the non-linear interaction of all variables and the global effect of researching Energy...

At bio expensive TT 20/25/30 is 246k / 666k / 1043k in a tiny universe, while e.g. death stars are 'only' 70k with con cheap. So more than 15k is pretty much ultra-long-term. The cost of TT is about the same as energy cheap -> exp, which gives a minimum penalty of around 20% (until tech 26), so I guess the conclusion is that it is not worth it.

The only open question imho is whether paying for popgrowth/hab with expensive or normal energy might be worth it. I guess a bi-immune AR with expensive energy could be interesting?

Report message to a moderator

Re: Energy at normal / expensive Sat, 07 June 2014 15:37 Go to previous messageGo to next message
skoormit is currently offline skoormit

 
Lieutenant

Messages: 665
Registered: July 2008
Location: Alabama
The more important benefit of TT is that it reduces your tforming costs by 30%. The TT techs themselves will only matter in the very late game. If you live that long, you likely will win anyway.

The analysis to be made is the effect of 30% cheaper tforming on your early econ growth. My intuition tells me that it makes a bigger difference for no-immune AR, since they will have more low-val planets, and those planets benefit proportionally more from terraforming.



What we need's a few good taters.

Report message to a moderator

Re: Energy at normal / expensive Sat, 07 June 2014 15:51 Go to previous message
mrvan is currently offline mrvan

 
Officer Cadet 1st Year

Messages: 220
Registered: May 2014
Sorry if I was unclear, that is exactly what I tried to say.

You get 30% reduced TT cost, but the ~150 points are equal to energy expensive -> cheap, which is a 20% overall boost until tech 26, and even more initially, so that's a bad trade.

After getting immunity/ies, AR doesn't have a lot of point mines left, so alternatives are SRT's (something like NRSE+NAS. -IFE+LSP, but chances are that you already used either NRSE or NAS to get points), or ~2% pop growth or almost a third of the divisor 10->25. Going 10->25 without 2nd immunity sounds like suicide to me, and TT plus 2i also sounds crazy because you will terraform at most 15 clicks per world, so the actual discount is max 450 resources per world (until you invest 250k in bio)

So, the only option left is a no-immune TT AR, which arguably has more synergy, but it also means that the terraforming starts from a lower level which means that the cheaper terraforming is negated by less resources. The immunity costs something like 300 to 400 points (assuming equal 1/x hab), so that buys you something like TT plus 2% growth and a minor SRT (or you can buy TT plus bio cheap, if you really feel like researching 4 fields instead of 3...)


[Updated on: Sat, 07 June 2014 15:53]

Report message to a moderator

Previous Topic: Tri Imune AR?
Next Topic: AR has no resources?
Goto Forum:
  


Current Time: Fri Mar 29 09:30:29 EDT 2024