Farming & Growth Meter Effects (for v0.3 ?)

For what's not in 'Top Priority Game Design'. Post your ideas, visions, suggestions for the game, rules, modifications, etc.

Moderator: Oberlus

Message
Author
emrys
Creative Contributor
Posts: 226
Joined: Fri Oct 24, 2003 3:44 pm

#16 Post by emrys »

Geoff the Medio wrote: The real important strategic point is the difference between your production and another empire's. If you have 100 PP / turn, and they have 100 PP / turn, the difference is 0. If you produced 110 a turn, the +10 PP / turn gives you a huge advantage. The same applies if you're both at 1000 PP / turn or 10000. Assuming you have equal tech levels and strategic effects are neutral, the majority of your PP's "cancel", so however much extra you have is your advantage. The +10 has the same value in all cases. You don't need to go to +100 or +1000 to get the same relative value when at 1000 / turn or 10000 / turn. (Ok, I admit this isn't that clean cut, as minor fluctuations would be more likely to wash away the +10 if you're making 10000/turn, but the basic concept is sound)
The basic concept is clearly NOT sound. If I take a fleet worth 110PP into battle against one worth 100RP, and strategic effects are neutral, my back of a fag packet estimate would be that I have a reasonably good chance of winning, after all I supposedly have a 10% advantage. If I take a 10,010PP fleet into battle with one which cost 10,000Pp, my fag packet estimate would be that it was anyone's guess... thus that +10PP/turn is not worth as much to me as it would have been before.
For food, if you produce 9 / turn and need 10 / turn, then +1 is all you need. If you produce 99 / turn and need 100 turn, you only still only need +1 more. Getting +10 more is no extra help. The player "hungers" for each additional food the same amount.
Well, getting +10 more allows me to pursue a strategy that will result in me needing 109/ food a turn, i.e. I can go on a colonising spree and support a tenth of my empire as developing worlds, just as the +1 from 9 to 10 allowed me to colonise a single planet when I had ten planets. The impact on my strategy is the same.

Ellestar's argument is much more cogent. The drive towards double focus and specialisation is much stronger under an exponential system.

Of course a lot of this discussion is now looking irrelevant, since we seem to be headed away from a fixed 0-10 range towards perhaps -50 to 50 or 0-100 or some such, and a larger range of possible meter values gives more opportunity to accomodate the range of effects by varying the relative bonusses (e.g. late game techs can give +10 rather than +1, so the range balancing doesn't need to be done in the mapping table.

This was the main source of the problem, since with only ten steps between min and max, there just wasn't much room to maneuver.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#17 Post by Geoff the Medio »

Ellestar's main pro-exponential argument was unrelated to my own, so comparing their cogency (?) is not appropriate. It was also basically irrelivant, since exponential meters are not necessary to make specialized planets more attractive. drek's current plan is give have techs and buildings give bonuses and benefits to planets under certain specialized focus, but not other planets. Thus you end up with specialized focus being much more attractive (at least at later times in the game).

Furthermore, Ellestar's conclusion actually supported me:
Ellestar wrote:I think that linear bonuses from meters are better.
Here's a few other ways of looking at things:

Why should a linear meter increase get the same proportional bonus? Wouldn't it make more sense to give a larger percent bonus to an empire with a larger percent higher meter? That is, give larger bonuses for more advanced tech, rather than relying on the meter to give the increases in an less controlled way.

If empire 1 has 5 and empire 2 has 6 in a meter, why shouldn't the situation be much better for empire 1 if it has 40 and empire 2 has 41? The latter case seems like it should be much more evenly matched than the former.

Why shouldn't an empire with a lead have to work harder and harder to maintain the same proportional lead as the values of the meters starts to grow? With expoential, the empire with the bonus ends up in a runaway lead, where even matching relative bonuses from then on by a lesser empire don't even things out. This is an aspect of the diminishing returns Ellestar referred to.

Also, keep in mind the fact that the game is more than one turn long. If empire 1 has 1000 PP, and empire 2 has 1010 PP, after 10 turns, empire 2 has 100 PP more. That is a significant difference already.

Also again, keep in mind that there is more than one planet. If an empire is making 10000 PP, it's probably got more than one planet... say 100. Giving +10 PP production across the empire, on average, gives a 1000 PP bonus, which is not insignificant.

emrys
Creative Contributor
Posts: 226
Joined: Fri Oct 24, 2003 3:44 pm

#18 Post by emrys »

Geoff the Medio wrote:Ellestar's main pro-exponential argument was unrelated to my own, so comparing their cogency (?) is not appropriate. It was also basically irrelivant, since exponential meters are not necessary to make specialized planets more attractive.
I read Ellestar's post as being in favour of a roughly linear mapping factor, and against exponential, since He felt that exponential bonuses would excesively encourage double focussing, rather than that they were necessary in order to encourage it. In other words it was a parallel supportive argument to yours. As such a comment that Ellestars argument was more cogent, and hence more convincing and in more need of a solution seems a sensible thing to say.
If empire 1 has 5 and empire 2 has 6 in a meter, why shouldn't the situation be much better for empire 1 if it has 40 and empire 2 has 41? The latter case seems like it should be much more evenly matched than the former.
Leaving apart the basic point that the discussion was about which mapped +1 to the same strategic value at all levels, and here you are stating clearly that the +1 from 40 to 41 would be worth less to a player than the +1 from 5 to 6:

Of course, as I stated if we have sufficient range of values it would almost certainly be best to embody the value of the object in the size of the bonus. The problem comes when we map that argument to the ranage we were intially discussing i.e. 0-10. Given so little room to maneuver, then either we would end up with early tech bonuses being +0.1 which seems to negate the point of a 0-10 range, or we have a very limited range of relative tech bonuses (i.e. the best uber tech level in the game can only be at most ten times better than the worst, and in practice probably less since you may want to have one or two bonuses that stack by that point)
Also again, keep in mind that there is more than one planet. If an empire is making 10000 PP, it's probably got more than one planet... say 100. Giving +10 PP production across the empire, on average, gives a 1000 PP bonus, which is not insignificant.
maybe I'm missing the point you were driving at here. It's just as significant as +10PP on each planet no matter how many planets are in the empire..., it's still +10%. The number of planets are irrelevant, the output of each (i.e. the total number of bonuses a player has collected) is relevant. Under a linear scheme the more bonuses you have applied, the less difference each additional +1 makes to you, under an exponential scheme the relative extra value of any tech that gives +1 is the same.

The only real differenc between the schemes is this, in order to have later techs be better in a linear scheme, they must increase rapidly in bonus size (at about the same rate as the typical economy would over the typical time to research them) and having or not having, stacking with or not stacking with earlier techs will become irrelevant rapidly. Under a geometric scheme, they should be of constant bonus size but stack with earlier ones, or include the bonus of the earlier ones if they do not stack. That's about it really.

Since I assumed we wanted later techs to be better than earlier(or at least similar), then the limitations of the 0-10 range semed to cause difficulties given the requirements on bonus growth that linear mapping gives, so I suggested geometric as a baseline, since it would leave us with ten distinct, equally significant phases for tech.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#19 Post by Geoff the Medio »

emrys: re: your reading of ellestar's position: oohh... ok. I understand what you meant now. Not having a good day for interpreting things not completely spelled out for me, apparently.
If empire 1 has 5 and empire 2 has 6 in a meter, why shouldn't the situation be much better for empire 1 if it has 40 and empire 2 has 41? The latter case seems like it should be much more evenly matched than the former.
Leaving apart the basic point that the discussion was about which mapped +1 to the same strategic value at all levels, and here you are stating clearly that the +1 from 40 to 41 would be worth less to a player than the +1 from 5 to 6:
"strategic value" is ill-defined. What you consider always equal, +25%, I consider wildly varying with current level.
The problem comes when we map that argument to the ranage we were intially discussing i.e. 0-10. Given so little room to maneuver, then either we would end up with early tech bonuses being +0.1 which seems to negate the point of a 0-10 range, or we have a very limited range of relative tech bonuses
Which is why we went to 0-100 or so, with +1 as the smallest bonuses. Does this mean you've fine with linear bonuses if we have a larger, and thus less granular, range of values?
Also again, keep in mind that there is more than one planet. If an empire is making 10000 PP, it's probably got more than one planet... say 100. Giving +10 PP production across the empire, on average, gives a 1000 PP bonus, which is not insignificant.
maybe I'm missing the point you were driving at here. It's just as significant as +10PP on each planet no matter how many planets are in the empire..., it's still +10%. The number of planets are irrelevant, the output of each (i.e. the total number of bonuses a player has collected) is relevant. Under a linear scheme the more bonuses you have applied, the less difference each additional +1 makes to you, under an exponential scheme the relative extra value of any tech that gives +1 is the same.
The main idea there was concerning your point about +25 PP being irrelivant when you've got a million PP. I was attempting to point out that the +25 would be only from a single planet, in which case a single increment of the smallest size available should be insignficant compared to the power of the force, er, the combined out put of the 1000's of worlds which would be needed to generate millions of PP / turn. If you took that +1 and spread it across all those worlds, then you'd get +25000 or so, which is still not as big a deal compared to a million as it would be compared to 100000 or so, but it's not nearly as bad as your 25 vs. 1000000 example might seem.
The only real differenc between the schemes is this, in order to have later techs be better in a linear scheme, they must increase rapidly in bonus size
This isn't entirely true. Latter techs already have the advantage that they add even more of a bonus. I realize you won't see this as making them "better", though.
Under a geometric scheme, they should be of constant bonus size but stack with earlier ones, or include the bonus of the earlier ones if they do not stack. That's about it really.
Hmm... I'm not sure what you mean here. A you saying that you're worried that latter techs won't stack with earlier techs? Meaning after you get +1 early on, the next tech would have to be at least +2 to have any effect? This isn't what I'm suggesting at all... The first tech would give you +1 and the next would give you an additional +2, so you'd end up at +3 total, and so on.

Alternative, do you mean that each tech gives +1 more than the previous one, so the first +1 gives +1, the second gives +1 +1 = +2 for a total bonus of +3, the third gives +2 +1 = +3 for a total bonus of +6? Technically, this would be a quadratic bonus, not linear... though it could also be emulated with a linear system just by making the first tech +1, the second +2, and the third +3 with the effects stacking (which is roughly what I expect to see implimented)
Since I assumed we wanted later techs to be better than earlier(or at least similar), then the limitations of the 0-10 range semed to cause difficulties given the requirements on bonus growth that linear mapping gives, so I suggested geometric as a baseline, since it would leave us with ten distinct, equally significant phases for tech.
I had no idea that was the basis of your position... The original main point seemed to be about making each +1 worth the same, about how to achive which my view is discussed previously. The current plan is to use 0-100 as the rough scale of the resource meters. The 100 could go higher / lower as needed. (I didn't think we needed a cap at all, but drek seemed to think we should have one...)

drek
Designer Emeritus
Posts: 935
Joined: Thu Jun 26, 2003 8:07 am

#20 Post by drek »

If you have a competing idea for the following numbers and system, then spell it out completly (as Geoff has done on the wiki talk page.) Unless most of us can point at a system and say "that'll obviously work" Aq might have to go set up a public review (an undesirable thing at this stage, since we are pressed for time.) or point at a system and say it's in (an undesirable thing cause it might cheese people off).

We do need to balance food production (and later mineral vs. industry production). It's a good thing that people are thinking about this sort of crap.

Assumptions:

* 1 unit of farming infra produces 1 unit of food per population.
* Primary foci adds +15 to Farming, Secondary +5
* There will be techs and buildings that improve the effects of Primary Farming, and Farming empire-wide.
* 20 is the breakpoint. You need 20 farming to have enough food for decent health, you need 20 Health to avoid population reduction.
* The homeworld gains +20 to Farming, because it needs to support undeveloped colonies in the early part of the game.
* Health rating of a decent planet is 50.

Nutrient Distribution:
Each unit of population requires Nutrients in order to survive. One point of Nutrients is required for basic survival; two points allow normal health; four points grant good health.

Nutrient Distribution must occur after resource generation, but before population growth is checked.

Each planet has a Nutrient Rating similair to in function to the Enviroment Rating. The Nutrient Rating is determined by dividing Local Nutrients with Population:

* If Nutrient Rating is 4 or above, the planet is "Well-fed." Health and Happiness increase by 10.
* If Nutrient Rating is 2 or above (but less than 4), the planet is "Adequately fed". Health and Happiness are unchanged.
* If Nutrient Rating is 1 or above (but less than 2), the planet is "Poorly fed." Health and Happiness are reduced by 10.
* If Nutrient Rating is less than 1, then the planet is "Starving." Health and Happiness are reduced by 50. In addition, Health is capped at 20.

During the resource generation phase of a turn, planets with more Nutrients than twice their Current Population export these excess Nutrients to the global stockpile.

During the Nutrient Distribution phase, planets that are "Starving" can import Nutrients from this stockpile to make up for shortfalls. In the first pass, planets import just enough food to reach the "Poorly fed" rating. If, after this first pass there are still Nutrients remaining in the global stockpile, then a second pass is run. During this second pass, planets import just enough food to reach the "Adequate" rating. A third pass is then run, with planets importing just enough food to reach the "Well-fed" rating.

Planets import food in order of colony age. The oldest worlds get first dibs.

Future note: Planets that are "blockcaded" cannot export or import resources.

emrys
Creative Contributor
Posts: 226
Joined: Fri Oct 24, 2003 3:44 pm

#21 Post by emrys »

Look, Geoff, this is really silly:

Imagine empire A and empire B, each of who have the same number of planets, each of which produce 10PP/turn.

empire A has the opportunity to develop a tech that adds +10PP/turn/planet. how much do they want this tech? well, it'll mean they suddenly become twice as powerful as their opponent, so they'd kill their gran to get it :)

empire A develops this tech, empire B also develops it. Now both empire's planets each produce 20PP/turn.

Empire A now has the opportunity to develop another tech, which gives another +10PP/turn/planet. How much do they want this tech? Well, it means they suddenly become about half as much more powerful than their opponent, so they want it about half as much. For them to desire the tech as much, it would need to add +20PP/turn.

What does this mean? Well, if the smallest bonus you can get under our 0-100 system is +1, then there are at most six fully stackable techs with equally desireable bonuses (i.e. +1,+2,+4,+8,+16,+32). Since that seems like it's probably going to be enough, I no longer see that we need to have a roughly geometric mapping, we can do it by choosing the bonuses themselves, when we were talking about integers 0-10, then there were at most 3, which didn't seem enough.

emrys
Creative Contributor
Posts: 226
Joined: Fri Oct 24, 2003 3:44 pm

#22 Post by emrys »

drek wrote:* 1 unit of farming infra produces 1 unit of food per population.
Seems to be a factor of ten missing somewhere:

1 unit of farming infra produces 0.1 units of food per population perhaps?

Also, any plans for an effect of Environment under this system? I'd like to see poor (orginal) environment planets produce less food, so either requiring good environment planets to shift to food production as you colonise more marginal worlds, or requiring you to build up your food tech to support those marginal worlds, or a bit of both.

drek
Designer Emeritus
Posts: 935
Joined: Thu Jun 26, 2003 8:07 am

#23 Post by drek »

1 unit of farming infra produces 0.1 units of food per population perhaps?
Correct. 10 units of farming infra produces 1 unit of food per population.
Also, any plans for an effect of Environment under this system? I'd like to see poor (orginal) environment planets produce less food, so either requiring good environment planets to shift to food production as you colonise more marginal worlds, or requiring you to build up your food tech to support those marginal worlds, or a bit of both.
Correct. See the Farming meter on the v.3 scratch pad:

http://www.freeorion.org/wiki/index.php ... ScratchPad

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#24 Post by Geoff the Medio »

emrys wrote:Look, Geoff, this is really silly
I concur.

In the Tour de France, you don't need to win by 50% more than your nearest opponent in every stage. You win by small linear amounts in each stage, which add up to a big difference at the end. Production is like a bike race.

I don't need to be 50% more powerful than the next empire... I just need to have an advantage in production every turn, which grows and grows over time, until the total advantage I've accumulated is enough to have a decisive power advantage.

If you have an advantage of +1 over an empire with 50 production, that advantage should be proportionally much less than if you had an advantage of +1 over an empire with 4 production. In the former case, it should and does take much longer to build up a sizable lead in total PP spent/accumulated, which is good. Matching gains over time should eventually reduce the importance of a starting bonus, so if you have +3 at the start o the game, and both empires get +50, the +3 is not significant in comparison to what you've developed since.
Empire A now has the opportunity to develop another tech, which gives another +10PP/turn/planet. How much do they want this tech? Well, it means they suddenly become about half as much more powerful than their opponent, so they want it about half as much. For them to desire the tech as much, it would need to add +20PP/turn.
Empire A now has the opportunity to develop another tech, which gives another +10PP/turn/planet. How much do they want this tech? Well, it means will get twice the bonus that they got with the previous tech, so they want it about twice as much. For them to desire the tech as much, it would need to add the same +10PP/turn. Furthermore, if Empire B hasn't gotten the first tech before A has the chance to get the second, the advantage is the same by both measures, whereas a growing % bonus would not be.
Well, if the smallest bonus you can get under our 0-100 system is +1, then there are at most six fully stackable techs with equally desireable bonuses (i.e. +1,+2,+4,+8,+16,+32).
I'm not sure what you mean by this... There's no reason we can't ahve 8 different techs and buildings each giving stacking +1's... (?)

Daveybaby
Small Juggernaut
Posts: 724
Joined: Mon Sep 01, 2003 11:07 am
Location: Hastings, UK

#25 Post by Daveybaby »

Edit : ooops - double post :oops:
Last edited by Daveybaby on Thu Jul 15, 2004 8:59 am, edited 1 time in total.
The COW Project : You have a spy in your midst.

Daveybaby
Small Juggernaut
Posts: 724
Joined: Mon Sep 01, 2003 11:07 am
Location: Hastings, UK

#26 Post by Daveybaby »

Geoff the Medio wrote:
Well, if the smallest bonus you can get under our 0-100 system is +1, then there are at most six fully stackable techs with equally desireable bonuses (i.e. +1,+2,+4,+8,+16,+32).
I'm not sure what you mean by this... There's no reason we can't ahve 8 different techs and buildings each giving stacking +1's... (?)
Yes there is! Its the whole point of his argument! If each +1 only gives a linear increase, then you arent getting a big enough increase from later techs to justify the effort.

Okay, lets look at this another way. Say you have a race which has a bonus to mining, so it starts with +1 mining on its meter. Then you have another race which doesnt have the mining bonus. By the end of the game the 1st race's mining bonus has become negligible - in fact every race has become virtually identical in performance giver or take half a percent here or there. I dont consider this a good thing.

And the tour de france analogy isnt really valid. But while we're making bad bike riding analogies :wink: If you are 25% percent fitter than your opponent, you expect to be 25% faster on the first stage, 25% faster on the last stage, and 25% faster overall. You dont expect to be 25% faster on the first stage, 12.5% faster on the second, 6.25% on the third, and 0.01% faster on the final stage.
The COW Project : You have a spy in your midst.

Impaler
Creative Contributor
Posts: 1060
Joined: Sun Jun 29, 2003 12:40 am
Location: Tucson, Arizona USA

#27 Post by Impaler »

We need to keep in mind that ALL bonuses are in a sense Multiplicative even IF they are linear.

For example lets say I make 5 units of money and my oponent makes 4. I am only making 20% more at first but I think INVEST that extra and build some additional infastructure or military or what not. Soon I have excelerated my production and am making 6, 7, 8 and so on. Any thing that kicks up the speed of my expodential growth will grow in importance over time and ultimatly become a BIG exponential advantage over time.

On a similar facet, consider how different bonuses synergize. Say for example a bonus that gives me more population and a bonus that makes my population produce more. When you combine these together the effect is not simply linear its exponential. A whole lot of little +1's can add up quite fast.

In the end though I dont have anything fundamentaly against the larger 0-100 scales your proposing, my consernes are mostly astetic as large numbers can feel "clumsy" some times compared to crisp and straitforward small numbers like thouse used on the SMAC social Enginering scales. I tend to be conservative here and stick with anything SMAC like.
Fear is the Mind Killer - Frank Herbert -Dune

emrys
Creative Contributor
Posts: 226
Joined: Fri Oct 24, 2003 3:44 pm

#28 Post by emrys »

Geoff the Medio wrote:
Well, if the smallest bonus you can get under our 0-100 system is +1, then there are at most six fully stackable techs with equally desireable bonuses (i.e. +1,+2,+4,+8,+16,+32).
I'm not sure what you mean by this... There's no reason we can't ahve 8 different techs and buildings each giving stacking +1's... (?)
If I produce 1 unit per turn, I gain a tech that doubles my output (i.e. has a+1 bonus), and I now produce 2 per turn. To double my output I now need a tech that stacks a +2 bonus on top, at which point I produce 4 per turn, and need a +4 bonus tech to double my output, etc.... Seems simple enough.

Now of course I'm not arguing that the (often run away) cumulative effect of bonuses isn't vitally important, because it clearly is. I'm just saying that prima facie to get about the same magnitude of strategic effect from a tech on a base level of 30 requires a much bigger bonus than to get that effect on a base level of 4.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#29 Post by Geoff the Medio »

Daveybaby wrote:There's no reason we can't ahve 8 different techs and buildings each giving stacking +1's... (?)
Yes there is! Its the whole point of his argument! If each +1 only gives a linear increase, then you arent getting a big enough increase from later techs to justify the effort.[/quote] Uhm... so?

Though I don't agree, for the sake of argument, let's say that +1's get less strategically useful for each one you get. That's still not a reason to outlaw having more than one tech / other thing give a bonus of +1. Not every tech has to give +1, but a given bonus size doesn't have to be exclusive to a single bonus "source".
Say you have a race which has a bonus to mining, so it starts with +1 mining on its meter. [...] By the end of the game the 1st race's mining bonus has become negligible
A race bonus doesn't have to be in the form +X to a meter. It could always add 20% to the meter, or to production per meter effect for the race: instead of

(production) = (pop) * (meter) / 10

A race could have

(production) = (pop) * (meter) / 8

for a certain resource.

Alternatively, a race could get +2 when other races get +1 from bonuses to a particular meter.
And the tour de france analogy isnt really valid. [...] If you are 25% percent fitter than your opponent, you expect to be 25% [on each stage] You dont expect to be 25% faster on the first stage, 12.5% faster on the second, 6.25% on the third, and 0.01% faster on the final stage.
You misunderstood the analogy. Each "tech" is not a stage. Each turn of the game is a stage. The % you finish faster than someone else isn't important. The time you finish ahead is important. ie. I finish 8 seconds a head, and I'm ahead by 8 seconds . Whether the stage took 9 years or 50 seconds, I'm still ahead by 8 seconds. If we both improve our time by 20 seconds, I'm still 8 seconds ahead.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#30 Post by Geoff the Medio »

Impaler wrote:In the end though I dont have anything fundamentaly against the larger 0-100 scales your proposing, my consernes are mostly astetic as large numbers can feel "clumsy" some times compared to crisp and straitforward small numbers like thouse used on the SMAC social Enginering scales. I tend to be conservative here and stick with anything SMAC like.
Originally (as drek conceived them, or perhaps later), meters were 0 to 10, in steps of 1. This was deemed to granular, with steps too big to have enough different meter settings. Thus 0 to 10 with steps of 0.5, 0.25 or 0.1 were proposed. I thought having +1 be the smallest bonus visible to the player was better, however, so I suggested 0-255 with steps of 1. drek then switched to 0-100 because 100 is a round number, which he likes.

Post Reply