A Quick idea Re research;

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

Moderator: Oberlus

Message
Author
User avatar
Robbie.Price
Space Kraken
Posts: 161
Joined: Sun Dec 16, 2007 7:00 pm

A Quick idea Re research;

#1 Post by Robbie.Price »

Goodmorning all,

I think it is highly unlikely that this will be used, but i wanted to suggest it anyway.

One thing that strikes me is that there are two very differnt types of progress in science, generally speaking.

There is the progressive continuous adaptation/application/tuning/and exploring of ideas/technologies. And then there is something completely differnt, the discovery of new ideas, new possibilities whole new fields of research being born.

One can think of a small number of ideas which are large steps away from previous works, which latter get refined and improved to produce technologies, or new ideas.

In general science has Large conceptual leaps, and small progressive improvements. I am hoping that we can somehow model these two differnt types of growth. (for a poor example the laser; the 'Concept' of energizing a gas between two mirrors and using stimulated emission to produce a coherent light source was really quite brilliant. . . . comparably the concept of making a very small laser, and a disk with very small mirrors and encoding information in the tilt of the mirrors 'A CD' isn't nearly as revolutionary 'doubly so since hard disks and record players already existed').

In the game some races could be better at the large conceptual leaps, or better at the progressive application(or both or neither).

Under this model the tech tree would be a lot of smaller techs, which are made researchable in large numbers by completing Very Very large research projects.

something like

Lasers, Las1, Las2, Las3 . . . .
Optical Storage Media
industrial cutting
. . .
Controlled projected plasma,
Plasma Cannon, PC1, PC2
. . .

To research Controlled Projected plasma one would need one or two levels in lasers, perhaps, and maybe one or two other tech in other fields. But one would almost never run out of 'things' which could be derived from lasers.

I'm not sure how to model it game wise. or how to balance it. . . but I wanted to put it out on the table to see what the FO community thought of the idea.

If people find it interesting perhaps we can find a way of incorporating it. . ., if the tech tree is not too set in stone.

Else it was just an idea.

Best wishes all

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12799
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: A Quick idea Re research;

#2 Post by Geoff the Medio »

FreeOrion already has three types of tech: Theories, Applications, and Refinements. Exactly how many applications per theory or refinements per application will depend on balancing and whatever content gets created.

User avatar
Robbie.Price
Space Kraken
Posts: 161
Joined: Sun Dec 16, 2007 7:00 pm

Re: A Quick idea Re research;

#3 Post by Robbie.Price »

That's sort-of what I thought there was, but does FO have any manner of making it so that some races are better at Theory techs, others better at application techs and still others excelling at refinements?

That's more the point I was going for, although re-reading what i wrote, i don't think I made that distinction as clear as I might have.

Best wishes all.

Price.

User avatar
Tortanick
Creative Contributor
Posts: 576
Joined: Sat May 26, 2007 8:05 pm

Re: A Quick idea Re research;

#4 Post by Tortanick »

We're not doing races for ages yet ;)

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12799
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: A Quick idea Re research;

#5 Post by Geoff the Medio »

There's no way now to change the cost or research time of techs (or buildings, or ship parts, etc.) other than changing the relevant tech, building, etc. definition in text files. We might add a way to do this sort of thing eventually, though.

User avatar
Robbie.Price
Space Kraken
Posts: 161
Joined: Sun Dec 16, 2007 7:00 pm

Re: A Quick idea Re research;

#6 Post by Robbie.Price »

Ok, I've been thinking about Tech; Here's some thoughts;

Motive: For me Theories, Applications, and Refinements are very different things, and I Hope that we'll be able to find a system where they can be treated adequately differently.



I understand there is no way to have each race require a differnt amount of RP for each tech . . . so i was thinking instead might it be possible to add another way to generate RP. . . .

When researching a tech you've got some number of RP you can put to it per turn (what you generate by colony settings). In addition we could have a set % of remaining RP's needed per turn completed. This set % would be what varies form race to race and from Theories vs Apps vs Refinements.

For example:

A race could be very good at theory, but normal at app techs and bad at Refinement techs

they would have something like:

10% of remaining needed RP points to complete a Theory tech would be given 'free' each turn (90% complete in 22 turns)

8% of remaining RP needed to complete App techs would be generated by the race (90% in ~28 turns)

but Refinement techs only 6% of the outstanding amount would be generated (90% in ~37 turns)

These are only guidelines ofcourse, For the curious the equation of how many turns to reach YX% given Z%/turn is: Turns =Ln(1-0.XY)/Ln(1-(0.01*Z)) where Ln is the natural Log button on a calc, or =ln( ) in Excel.

IF we add a system like this there are a LOT of things we can additionally do which would give the game's research a lot more depth, and be more fun, in my eyes.

For example:
1. When trading for a tech, or stealing a tech, or researching a tech from a prototype you scavenged from the battle field ... whatever . . . All the code would need to do is add a few %, to drastically speed up the process(depending on how you what type of aid your getting). (under diplomacy preliminary viewtopic.php?f=5&t=2040, trading tech's gradually was supported by 'most'? of the people who posted) this would be an additional way to implement the process

2. We could define a 'completeness' threshold, before which you can't being to apply colony generated RP to a tech. Theory Techs might have to be 70% complete before you can 'target' them, while Apps techs would only need to be 40% or 50% complete, and refinement techs 10%-20% complete. In science a lot of New ideas are proposed, most of which are junk, it's only after an idea has to some extent proven itself that governments/industries will take interest. New theoretical ways of looking at the world must be much more robust, defended before people are going to think about accepting them whereas building a better mousetrap . . . well you probably get the idea. Ofcourse this would mean as soon as you have the prerequisites for a tech it would automatically start to be researched by the population, and some number of turns later it would become 'unlocked' for the contribution of colony RP.

3. If you think about it, Since the % is applied to uncompleted research . . . it naturally puts an slight extra price tag on 'rushing' a technology. For example, a tech which is 'just' unlocked has 1000 more RP till it's done. If you invest 40 RP a turn each turn you'll finish it in 15 turns, and have invested ~594 RP(the rest being generated by the 'normal rate' of 6% of remaining per turn), however, if you wait 12 turns, then start investing 40 RP per turn, you'll finish the tech on the 21st turn, but only have invested 350 RP. It becomes an additional strategic decision when to start investing directly . . . also note that techs will almost never finish themselves . . . at some point some colony RP must be spent but the longer you wait the less you'll have to pay.

4. With this system we can also have a 'fluidity' to technology. In most 4X games one empire can invest heavily in Tech and if the game is somewhat unbalanced easily accumulate all, or most of the techs and be 'ages' beyond their opponents. we could set up a system were if your neighbor has a technology you can research, then you get some fraction of a % point to your natural rate. Additionally we could have it so the closer your two empires are, the more you trade, have general agreements the larger that fraction.

4.B. Some techs would be more liquid then others, I'm thinking mostly of Theory vs refinement techs, If you point your brand new just discovered Telescope into the sky and see the Hubble Space Telescope orbiting the planet. . . well at least your going to know that orbital mechanics is possible, you may not KNOW anything about it yet, but it'll get you thinking, and that's half the battle. Comparatively Refinement techs would not be very fluid at all. even though the theory of laser are the same everywhere the manufacturing processes and materials used to make them might differ from one race to the next. As such the fact that Race A can make level 7 lasers, and you can make only level 3 lasers won't help you much in researching level 4, since advances are so much more specific to how you make your lasers... if that makes sense.


I think on a code level it would be reasonably easy to code RP_needed = RP_needed*(1- 0.01*Natural_Rate) - Colony_RP; aught do it. Where Natural Rate would depend on how good the race is at that type of research (Theory,App,Refine) + some a small set of one time bonuses depending on weather or not the race is receiving any help from other races(legitimately or not).

I think I've written more then enough on the topic for now, does anybody else think this would be worth implementing in FO, or a similar system? I should hope it's enough to give a distinct 'feel' to each of the tech types, + provide some more dynamic strategic variability...

Anyway them's be my thoughts, what be yours?

Robbie Price.

User avatar
Krikkitone
Creative Contributor
Posts: 1540
Joined: Sat Sep 13, 2003 6:52 pm

Re: A Quick idea Re research;

#7 Post by Krikkitone »

Actually to change the "cost" for a race it might be simpler to give them a free bonus based on how much they put into the tech/theory/application

ie a race gets 0.1 free RPs in a Theory for every RP they put into it

which means that the Actual cost hasn't changed, but the Effective cost has changed.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: A Quick idea Re research;

#8 Post by eleazar »

There are some interesting ideas here, but the general impression i get is that you are trying to make a more accurate simulation of the process of scientific discovery, which of course isn't necessarily the same thing as game mechanics which are fun.

You need to explain why it may make the game better not more realistic.

M4lV
Space Squid
Posts: 89
Joined: Sat May 10, 2008 10:51 am

Re: A Quick idea Re research;

#9 Post by M4lV »

I agree here with eleazar, though that logarithmic formula is easy to understand for a math student even in his first semesters, you will scare others off that don't have a clue how to deal with it.

I'd make it a lot more linear and simpler. The remaining RP idea is good enough for that. We at our game project have solved the idea as follows: Allocated RPs are 1:1 used and added up each turn for tech development, no race penalties there (the only penalty would be different RP production of their Research Centers), but there is always a race-dependant chance when a tech is over 66% completed, that all(!) the remaining RP are generated for free and the tech is yours. There lies the possibility to distinguish between applied sciences, theoretical sciences and refinement research. Each tech can be classified as belonging to one of the three and then each race can have a clear attribute calling it good/medium or bad at each of the three influencing the chance to get a tech for free once having developed it far enough.

On a side note, even your gas and laser example wasn't achieved out of nothing. There was Planck and Einstein and Pauli before that laid groundworks for this experiment to even be able to be carried out. Also the necessary material must be developed and engineered first. Take a look at Lavoisier and his water separation experiments. Without good money and perfectly balanced/gauged/calibrated measuring instruments, no scientific advancement other than "braintalking ones" could have been achieved by him.

So there's no logarithmic math involved (besides maybe for the random generator ;)) in all this and it's based on randomness, understandable for the player. If randomness is not wanted in-game, then it's better to let go of the whole feature and think about something else making Research a little more variable thus exciting. Letting a tech to be developed in parts on its own by the population is quite an unconventional idea for such types of games like FO is (but nevertheless an interesting one).
But we should make it less complicated for the (mostly too young) gamer. Maybe we could make this work only for one or two races at all. These races would then have this kind of Research development as their speciality and could be avoided by too inexperienced players easy enough.

User avatar
Robbie.Price
Space Kraken
Posts: 161
Joined: Sun Dec 16, 2007 7:00 pm

Re: A Quick idea Re research;

#10 Post by Robbie.Price »

Goodmorning all,

ok, Important things first

I'd like to thank eleazar for the polite objection to the emphasis on realism in the first post, upon re-reading it i realized that i had not adequately defended the game reasons, and i appreciates the politesse with which you pointed that out.

My Game motivation is divided into two major parts: strategic depth, and to a lesser extent immersion.

1. strategic depth:

With the linear system, the strategic depth of how you approach research is somewhat limited. You can choose how much of your empires time/energy is devoted to research, you can choose what order to do your research in, and you can trade techs for other things. Crucially there is very little feedback between these systems, the only exception being economy techs, and research techs (which give you more money or more research per unit money[directly or indirectly]). So to a greater or lesser extent how you approach research is largely irrelevant. In my proposed system, the order which you do research can to some extent change how many total RP you have to expend by the end of the game. If you trade away techs, you also gain a little in research rate*assuming they have techs you don't* since trading increases contact between nations which increases the fluidity of their techs into your empire. Even when you choose to start investing in a given tech becomes strategically interesting. Compared to a linear system, an exponetial system has a lot of depth, which generally -> more fun. Additionally the 'cost' it terms of explanation and coding is quite small. as mentioned above.

For a few more specific examples: Balancing cost vs finishing time, and general strategies of technological growth.

In a linear system if you want a technology sooner all you have to do is invest more/turn and you'll get it done faster (or trade for it/steal it). 'Rushing' a tech has no 'cost' associated with it (except the very basic level choosing one tech over another, choosing research over ship production . . . ). Thus one's strategic field is limited. With the exponential system, you gain an additional dimension of game play: resources invested in research earlier contribute less, so you can rush a project but the more you rush it the more the rushing costs.
In a exponential system patience is not only a virtue, it's an important strategic consideration.

Re: general Strategies:
With a linear research model, there are only very weak feedback links between discovery order and research output. The order you research technologies in, and when you invest in them can not effect the research process very much. For example, if you have a tech which will cause all researchers to produce +1 research, and a new engine both of which cost the same number of RP, you're better off researching the +1 first then the new engine, if you can wait for the engine. Doing the research research first gives you both techs sooner then getting the engine first. To a lesser extent similar things are also true of getting economy techs early, so they you have more money to spend on research . . . so on and so forth, but that's the real limit.
In my proposed exponential system, you could, if you wanted (and no enemies were going to destroy you), get all the techs in the game for 1 RP each *you'd just have to wait a VERY VERY VERY long time*. Additionally even the order you get your techs in matters. If you get a tech which is a pre-requisite for another tech, then research automatically begins on the new tech. Finishing techs, which many other techs depend on, early generates extra 'natural' research points/ turn earlier in the game. This gives an extra level of feedback. Additionally there is no reason why some techs couldn't be made 'loosely' prerequisite on another. Higher shield technologies helps greenhouse and farming technologies a small (tiny) amount, higher level Laser Cannon refinement techs helps future plasma cannon refinement techs since the problems are 'similar' and so forth. The more self-referential, the more opportunity for feedback the system the more dynamic, alive, strategically interesting (and fun) it will be IMHO.

I'm going to finish This post on Why,

and then post another on the How's replying to comments made by others.

Best wishes,
Robbie Price

User avatar
Robbie.Price
Space Kraken
Posts: 161
Joined: Sun Dec 16, 2007 7:00 pm

Re: A Quick idea Re research;

#11 Post by Robbie.Price »

Goodmorning all again,

I noticed i forgot to talk about Immersion, under game reasons. By creating a more dynamic system, and by linking that system more closely with other systems like Trade/ empire relations/ diplomacy/ etc you develop a more engrossing system. This would add immersion. *quick and to the point, how rare for me [unfortunately]*
To reply to some particular comments

@ Krikkitone:
Yes, this method would make the very basic desire to distinguish between theory - apps - refinement. But it lacks the interconnectedness I was additionally hoping for, but didn't properly express.

@ M4lV:

Re scaring off users: In short: As long as we make it transparent enough the underlining system shouldn't matter.

In long :
This could be a genuine concern. Although I can't bring myself to believe that any significant fraction of those who would want to play FO would be incapable of learning the basics of exponential functions and how to play with them, I can imaging many who don't want to have to bother and that's fair.
But consider, what does the user need/want to know?

When will this tech be done? how much RP will it cost? can i get it faster? and can i cut costs? and a more general, what is generally the best stratagem for research....

We already have a "turns left" indicator. So weather the underlining system is linear or exponential or a mix won't matter the user will always have a concrete number of turns left.

Cost: we could add a indicator "Cost Remaining: 600 / 1000" meaning with current spending you'll invest 600 rp to get the tech which would be 1000 RP if not for the exponential behavour, if you continue to invest as you are now, or just 600 remaining whichever we want.

Faster: The short answer, of course yes. I covered this under 'Game' reasons

Cutting Costs How UI:
To integrate a exponential system I would add two (max four) buttons, to the leftmost list of techs. See the research screen shot http://www.optisch-edel.de/fo/screenshots/research.jpg each of the listed techs, "Arch mono", "Orbital Infra", "Genome bank" ... I would add a "Wait one turn" "Wait one less turn". (wait 5 more turns, start now buttons would also be good to add). Clicking one of the two would re-calculated the finishing times of the effected tech this would give the user all the 'Control' and all the information they need to decide, strategically, weather they want to wait or not (by looking at how many RP's they save by waiting, and looking to see if the number of turns till completion changes [sometimes clicking 'wait one day'] would still have the tech finish on the same day if the natural rate is much lower then the colony RP rate). This could be made so that clicking 'wait one turn' calculated the maximum number of days which it could not invest RP to cause the finishing date to increase by one... if we wanted to be nice (and have the system auto re-invest RP if that turn's RP investment is 'irrelevant' if we wanted to get really fancy wouldn't be hard to self optimize).

As long as the user can See the effect of their choices, and plan accordingly then the underlying complexity adds depth without adding too much confusion.

And best practices: This also ties more into 'game reasons', so it's handled above as well.


Re Random Rolls: I would hope to avoid Random Rolls, even if the probabilities are race dependent, if it could be avoided i would hope to avoid it. Random does not mesh well with strategic typically.

Re lasers being dependent on previous techs: Your absolutely right That's why techs have prerequisites. My point *although realism based* was that when lasers were first being built, they were not accepted publicly or moved to being high priority research until after they were mostly finished being developed, + Lasers would be an Apps tech. General Relativity or the theory of Gravity would be Theory techs, they had pre-requisites as well, but nobody could have really said "Ok, we're going to invest 40 RP a turn, and solve this Gravity problem" it took a remarkable breakthrough. . . . not that those fact are important. Everythings got prerequisites. Technology is inherently iterative.

Re having only for one or two races:

While this might seem simple enough, balancing it would be VERY difficult. Exponential based techs would be balanced with larger differences between the RP needed between techs. Also it would be a lot more work, essentially building two research models rather then one.

I honestly believe we can make a exponetial which is sufficiently transparent that even the younger kids can get it. it's only scary because it sounds mathematical, if you look only at the UI then the underlying complexity isn't very differnt at all.

Hopefully i did a better job of explaining the system,

best wishes all

I welcome further comments/ideas.

Robbie

M4lV
Space Squid
Posts: 89
Joined: Sat May 10, 2008 10:51 am

Re: A Quick idea Re research;

#12 Post by M4lV »

how do you want to handle the extra-micromanagement, a strategically optimised tech-handling would undoubtedly require since at each new tech breakthrough, it's necessary to relocate/reduce RP production and incrementally increase it with each turn passing.

You have to consider that such an exponential system requires "exponential handling" and turn optimization on the user side, otherwise I as a strategist would feel that I'm wasting too much production if I don't. Now what we would need then would be a system that automatically adjusts production according to your system.

The basic thing is, most users, even me as a proclaimed experienced 4X game player, don't want to allocate Production points and worker distribution each turn anew but on significant discrete points, i.e. when your techs are developed to a certain extent and you want to reduce/increase production in order to get faster to your aimed goal. With random rolls, you got your penalties too in doing that, the only thing is that they depend on randomness. Purists would deny such a thing, but making it exponential gets hard on the purists too or can you think about a distinct, easy-to-use-and-configure system to make strategic use of these new options without having to do it manually for all planets each turn?

That'd be the challenge in my eyes. Otherwise your idea is well-thought-out if one accepts the fact then, that research is done without allocation of RP of any kind, which, again, might seem strange. Does that "hidden" research or better the output of that hidden research depend on global production values or the number of your planets/population or anything? Or is it just the race's characteristics that sets a specific ground percentage that never changes resp. can't ever be changed by current game status?

User avatar
Robbie.Price
Space Kraken
Posts: 161
Joined: Sun Dec 16, 2007 7:00 pm

Re: A Quick idea Re research;

#13 Post by Robbie.Price »

M4lV wrote:how do you want to handle the extra-micromanagement, a strategically optimised tech-handling would undoubtedly require since at each new tech breakthrough, it's necessary to relocate/reduce RP production and incrementally increase it with each turn passing.

You have to consider that such an exponential system requires "exponential handling" and turn optimization on the user side, otherwise I as a strategist would feel that I'm wasting too much production if I don't.

Evidently my first attempt to outlined a method for avoiding such micromanagement was inadequate so is hall try to make it more clear.

At any given turn, you can always see the 'completeness' and 'turns till completion' of every tech your are researching.
Due to the nature of exponential functions you would either want to invest nothing, or a significant quantity of research into a tech. There is virtually no benefit to turn by turn increased allocation [because the system is exponetial, the more you try to micromanage the less important the additional micromanagement is. I could go into detail why, but it would be out of the scope]. Typically, when a new tech becomes available to be researched, The near maximum reasonable micromanagement would be "Turn 1: Pause the Research for N turns, [ensure the focus is reasonable on that tech], Turn N+Y(Y being ~half the turns remaining after N till 'finished') Increases focus on the tech to finish it off." Adding a third step, would 'gain/save' you a number of RP much much less then pausing for one more turn in the original N turns), Adding a fourth step would be trivial and any more then that. . . and well hopefully it's calming to your OCD :- ), because it's not Gaining you anything in actual game play.

Essentially from the time you start investing colony RP to a project you can safely treat it as almost linear(since the exponetial component is less significant each turn, therefor the system approaches linearity).
M4lV wrote: The basic thing is, most users, even me as a proclaimed experienced 4X game player, don't want to allocate Production points and worker distribution each turn anew but on significant discrete points, i.e. when your techs are developed to a certain extent and you want to reduce/increase production in order to get faster to your aimed goal. With random rolls, you got your penalties too in doing that, the only thing is that they depend on randomness. Purists would deny such a thing, but making it exponential gets hard on the purists too or can you think about a distinct, easy-to-use-and-configure system to make strategic use of these new options without having to do it manually for all planets each turn?


In an exponential system there is even Less reason to do planet by planet anything then in a linear system. Either the Exponential component is dominant, and nothing you do with your colonies matters enough to bother, or the linear system is dominant, in which case it's linear. . . and again you have no MORE reason to micromanage then with a linear system, in fact you still have less since the system is slightly exponetial. As I suggested above, for any given tech even if your empire remained static for the entire duration of the research, changing the settings more then twice would be pointless (you're likely to change the focus on the order of twice for empire driven reasons anyway).
M4lV wrote: That'd be the challenge in my eyes. Otherwise your idea is well-thought-out if one accepts the fact then, that research is done without allocation of RP of any kind, which, again, might seem strange. Does that "hidden" research or better the output of that hidden research depend on global production values or the number of your planets/population or anything? Or is it just the race's characteristics that sets a specific ground percentage that never changes resp. can't ever be changed by current game status?
The % 'natural research' as i like to think of it, (Colony RP being 'Government spending research' or 'directed research'), would by definition be natural, and the global production values/pops/ ect would not be taken into account.

However it wouldn't be strictly static either:
I mentioned that techs would have some form of 'fluidity' across political boarders, So if you have a lot of general trade and political agreements, and understand well the language of an neighboring empire(there is talk about making every other active empire's primary language a research branch), AND they have the technology your currently researching you would get a 'small' bonus to your 'natural research %' since your scientist would be exposed to functional versions of the thing your trying to build, or theory your working on [refinements would be much less fluid, and this effect would be more negligible].
Also if you signed a research agreement with a neighboring empire, you both would gain some what on your natural rates (differnt ideas from working with scientists of a differnt culture/way of thinking).
If you traded for a tech (another thread suggested tech trades should not be instantaneous)(or stole it) then you would get a significant bonus to the natural rate for that tech.

But all these effects would be Galactic policy decisions, and thereby increase immersion. *making a mineral for food trade agreement, not only effects your food, and minerals, and trust shared between your empires, and what your empires peoples think of each other, But it additionally would effect the fluidity of technology between you. The game is more responsive and interconnected . . . more immersive*

RE: "Now what we would need then would be a system that automatically adjusts production according to your system."

Here my hope is that the system will not be nearly complicated enough to merit any but the most superficial of automatic adjustments. The general feeling of the FO community, if i understand it, is "If a system is complex enough to make the user wish it was automated it's far to complex". That being said, I believe that my proposed system is nowhere near complex enough to merit any automation. If we wanted to add any automation having the system automatically not invest colony RP on turns where the investment would not reduce the 'turns left' counter. IE, if next turn this tech will have the same number of turns left as this turn, then don't invest Colony RP in this tech this turn.


As a last comment:

When I first suggested this system, I suggested a fast rate of 10%/turn, a medium rate of 8% per turn, and a slow rate of 6% per turn.

I later suggested that techs could have 'weak' prerequisites on each other. . . as well as other methods of boosting the "natural rate". As such, if we include weak prerequisites, and other such systems 10 8 and 6 would be too strong. I would now suggest closer to 6, 5 and 4%. The differences between the old system and the new would be re-covered by the immersion elements of exponetial systems.

For explination of weak prerequisites:
If a given theory tech has another two theory techs as prerequisites, then the application technologies of the two prerequisite techs would each provide ~0.5% to the natural rate of the new theory tech your now working on. Additionally if any of the applications have refinements, then those refinements would give each give progressively less bonus to the natural rate of the new theory (~0.2%, ~0.1%, ~0.05%...[Math majors will notice this is divergent... don't worry it's capped by completion of the tech in question]). Other *meshes* could also be added, depending on the complexity we're willing to add, and the amount of time we want to spend balancing. Or we could stick to 10 8 and 6 of course >:- D


Best wishes all,

Robbie Price

M4lV
Space Squid
Posts: 89
Joined: Sat May 10, 2008 10:51 am

Re: A Quick idea Re research;

#14 Post by M4lV »

As far as I understand research production, apart from the "for-free" one, is linear since the more percentage of workers or funds Research gets, the more output I gain. So when planets grow or tech upgrades take place each turn, my total RP output increases mostly linear (though planetary growth is exponential, but let's say, all planets have full population). Now each turn my RP increases linearly if left unmanaged but research is run with an exponential part in the formula. In order for your system to work with the rest of the game, all production must be logarithmic too. The efficiency curve is not linear since it's better to allocate the linearly growing production to other areas than RP where usage is strictly linear and efficiency does not depend on logarithmic functions.

In a linear approach I don't have to project my RP that much into the future to be optimal. You forget that the remaining turns for research to complete is not constant until the current tech project is finished. It varies with natural RP growth.

Besides, unlike with going linear, you'd need to take back all research production at the beginning of each new tech development. I don't need that with my linear approach when calculating simple completion percentage growth the right way. All surplus percentage can be allocated/distributed automatically to other tech development areas and projects so nothing is wasted. That means you can pretty much leave research management alone for quite some turns without losing much efficiency in the linear approach while yours require constant reallocation with each tech being developed.

Considering we have a lot of techs in FO and a huge galaxy, people might actually be glad not to be required to do that in order to keep up efficiency degrees with their fellow players but that remains to be seen how much the effect would be.

User avatar
Robbie.Price
Space Kraken
Posts: 161
Joined: Sun Dec 16, 2007 7:00 pm

Re: A Quick idea Re research;

#15 Post by Robbie.Price »

M4lV wrote:As far as I understand research production, apart from the "for-free" one, is linear since the more percentage of workers or funds Research gets, the more output I gain. So when planets grow or tech upgrades take place each turn, my total RP output increases mostly linear (though planetary growth is exponential, but let's say, all planets have full population). Now each turn my RP increases linearly if left unmanaged but research is run with an exponential part in the formula. In order for your system to work with the rest of the game, all production must be logarithmic too. The efficiency curve is not linear since it's better to allocate the linearly growing production to other areas than RP where usage is strictly linear and efficiency does not depend on logarithmic functions.
I'm not entirely sure what your trying to say at the beginning of this. What "For free" one??
note if your pop is maxed, and your planets are maxed, then your Research production is remains unchanged.

I'm not at all convinced that everything else will have to be exponential, really that doesn't make very much sense. RP are things you make, they have "cost(s)" in that your colonies can't be doing everything at once fully efficiently. . . Just because one of the systems has a decaying exponential term in it doesn't mean that all the other jobs have to have decay terms. At any given time we're researching between 3 - 7 things at once. . . of which 3-4 will be open for colony RP. By the time your choosing to distribute RP to a project the project is basically linear in behavior. . . the exponetial component only really matter before you un-pause research in that field.

also any question of 'efficiency' is slightly silly since it's just a balancing question don't you think. if you choose the natural rates well, and the daily colony RP production well, the game will be balanced and the 'efficiency' will be 'right'.

Also you NEED to do research, you can't win without doing some, so if it's not 'efficient' . . . "shucks"... maybe i'm not understand your objection could you word it more clearly for me why this could be a problem, I'm really not seeing it.
M4lV wrote: In a linear approach I don't have to project my RP that much into the future to be optimal. You forget that the remaining turns for research to complete is not constant until the current tech project is finished. It varies with natural RP growth.
I'm not really forgetting it, it's just that i don't see this as important... Weather the #RP available/turn grows linearly, exponentially, hyper-exponentially, or chaotically. . . as long as the system is balanced which is not really any harder then balancing a strictly linear system *well ok, hyper-exponential and Chaotic would be a bit harder. . . but the point is if you care you can make it better then you can ever make a linear system, no matter how long you try or how much you care.*
M4lV wrote: Besides, unlike with going linear, you'd need to take back all research production at the beginning of each new tech development. I don't need that with my linear approach when calculating simple completion percentage growth the right way. All surplus percentage can be allocated/distributed automatically to other tech development areas and projects so nothing is wasted. That means you can pretty much leave research management alone for quite some turns without losing much efficiency in the linear approach while yours require constant reallocation with each tech being developed.
I don't see an interface for focusing RP's more or less towards given tech projects. . . if such a system doesn't exist . . . then no worries.

However if a system for selective focus does exists, each tech can have multiple pre-requisites(the tech tree is a rather tangled web, at the theory level at least) . . . so one might safely ask from which pre-requisite does it draw it's 'focus' value from. The easiest system would be that by default, all techs start with a 1/# of techs being researched focus value. . . as such that's a perfectly reasonable level, and probably doesn't need adjusting (other then to pause it for N turns, if you decided you don't want to do rush the research).
M4lV wrote: Considering we have a lot of techs in FO and a huge galaxy, people might actually be glad not to be required to do that in order to keep up efficiency degrees with their fellow players but that remains to be seen how much the effect would be.
There you've lost me completely I really have no idea where 'efficiency degrees' is coming from... or what your trying to say by it. :?:

That being said, in short: the exponential system is really much simpler then your working it up to be.

In brief:

A tech has a balanced cost. Some turns after the pre-requisite are finished it becomes available for the user to chose to rush finish or not. the number of turns between pre-requite finishing, and being able to start rush-finishing a task depends on many factors(all balance-able): the race's natural ability, any neighbors having the tech, finishing other weak prerequisites, having stolen or traded for the tech, having generalized research pacts with other races, and of course game special.

After the tech become enabled for depositing RP the user then chooses to RUSH it to completion or not (with a cost [the cost is determined via an underlying system which is exponetial but who cares really what the underlying system is, only how it effects the user]), and by how much they want to rush it.

Now with the system I've suggested the 'default' behavior I've suggested is to 'Rush' everything, so in fact the active choice is to 'Wait Patently' for N turns, with an ,'exponentially determined', reward for patience.

The mathematically component is only a means to an end. Dynamic delays between pre-requisite finishing and next tech becoming available determined by a significant array of race and empire game state to increase immersion. Easy incorporation of the goal of having stolen and traded for technologies contribute gradually to research rather then just give you the final product all in one go. Arbitarally strong distinctions between Theory, apps, and refinement techs. Patience vrs Rush completion in technology growth. Probably others which I'm forgetting.

All of these things are "Possible" . . . maybe . . . with a linear system . . . but none of the linear solutions are as Easy to code and/or balance as with an exponential system.

Forget about the questions of efficiency, the mixing of linear and non linear systems, worries about people not understanding(because it's really not that complex) ... are the goals worth while? and am i mistaken in thinking that they can be achieved via exponetial decay? Can they be equally well, or better, achieved by some other quasi-linear, or even linear method?

Those are the important questions, everything else boils down to balancing (which we'll have to do anyway).

Best wishes

Robbie Price

Post Reply