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

Re: A Quick idea Re research;

#16 Post by M4lV »

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.
with "for free" I meant your idea of "unrushed" research progress.

If pop and planets are maxed, RP production still linearly increases with upgrades of your research centers/buildings/structures. And it varies linearly with worker allocation towards these buildings. I'm coming mainly from our game where we have everything regarding colony management already coded, so it could be a good source for closer investigation of the problem. So let's say I have a system (not necessarily the FO system later on) where pop equals available workers, i.e. system got 10 pop, I got 10 workers to distribute to certain areas/buildings/structures/mines/whatever. I choose to have 5 factory buildings that do my shipbuilding and upgrading things, 2 to energy production, 3 to research and 2 to food production. Now while my 3 researchers are working, my 5 factory workers can work on improving their conditions, i.e. upgrading the research labs/facilities and so on. So when that is finished at a given turn, I get more research from those 3 workers. What I say now is that if these upgrades give strictly linear research bonii, then overall, my RP increases linearly over the turns and such the optimal point of decision when to rush a project becomes difficult to calculate, at least more difficult as if the RP bonii are logarithmic. Then their effect would be less each upgrade and give me less variation to consider when to jump in and rush the project.

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

Re: A Quick idea Re research;

#17 Post by Robbie.Price »

M4lV wrote:
with "for free" I meant your idea of "unrushed" research progress.

If pop and planets are maxed, RP production still linearly increases with upgrades of your research centers/buildings/structures. And it varies linearly with worker allocation towards these buildings. I'm coming mainly from our game where we have everything regarding colony management already coded, so it could be a good source for closer investigation of the problem. So let's say I have a system (not necessarily the FO system later on) where pop equals available workers, i.e. system got 10 pop, I got 10 workers to distribute to certain areas/buildings/structures/mines/whatever. I choose to have 5 factory buildings that do my shipbuilding and upgrading things, 2 to energy production, 3 to research and 2 to food production. Now while my 3 researchers are working, my 5 factory workers can work on improving their conditions, i.e. upgrading the research labs/facilities and so on. So when that is finished at a given turn, I get more research from those 3 workers. What I say now is that if these upgrades give strictly linear research bonii, then overall, my RP increases linearly over the turns and such the optimal point of decision when to rush a project becomes difficult to calculate, at least more difficult as if the RP bonii are logarithmic. Then their effect would be less each upgrade and give me less variation to consider when to jump in and rush the project.
Infact if you've got "N%" bonuses to research, which produces a tech which increases productivity/cost by "M%", so that you can increase the number of researchers by :

Old_production * 0.01*M, which gives you (1+0.01*N)*Old_production*(0.01*M)

you no longer have a linear system, you have a Multi linear system. Which must be solved by matrix multiplication if you want to maximize both Production and research, with the additional complication that the Weighting factor for how much one RP is worth vrs one PP, to the User is at best highly multi-linear (and in fact is more emotional then logical in 90% of cases ) . . .
In short even for 'linear' systems(which are not linear, they are at best multi linear) 'calculating' the 'optimum point' is a joke. It simply can't be done meaningfully. Making the generated research exponential (And I'm not even sure how you would do this? RP_generated = A*B^C, where A is a constant, And B or C is the number of workers??? ) would in fact only make it harder. . . since each planet would have wildly differnt research/person efficiencies. . . or even more nutty effects. . .


How would you make a 'logarithmic' or 'exponetial' production anyway . . . I get the impression your not thinking through your arguments to their full conclusions and implementations. I could be wrong of course, but I'm certainly not understanding where your going.

Best wishes.

Robbie Price

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

Re: A Quick idea Re research;

#18 Post by M4lV »

A linear system does not have the problem to decide a point where to rush in cause there is no "for-free" development. I think you misunderstand me there. I'm all for "no-research-done-by-pop-for-free", not for "linear-research-.." while you are for "log-research-..".

In a system where RP are only generated by (even non-linearly) improvable research buildings, I don't have any problem, since it doesn't matter when I get more RPs, the tech projects are all finished the same way and the system of RP production is self-adjusting. Tech developments only mark waypoints on a clearly pre-determined curve.

It's the decision-making requirement that I'm complaining ;) about in your idea. I don't want to make any decision on each tech when to rush best, no matter if the required math behind it is multi-linear, logarithmic or non-standard analysis.

Especially since getting the optimal point of RP investment is a joke to calculate and largely depends on "pi times my thumb", I don't want it in the game.

Btw. I do think my suggestions through, as I can even prove ;). Just have a look at the botf2 projects and mine/ours in particular and see how I think it is easiest and still interesting/variable enough. Don't want to make commercials for it but if you're curious to know what I mean in depth, go have a look at it :). I don't know much about Space Empires or SotS or SoaSE or GalCiv, but much about botf (and botf2). If you happen to have ST:BotF, you can also take the idea from how it's handled there though it was less sophisticated there..

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

Re: A Quick idea Re research;

#19 Post by Robbie.Price »

M4lV wrote:A linear system does not have the problem to decide a point where to rush in cause there is no "for-free" development. I think you misunderstand me there. I'm all for "no-research-done-by-pop-for-free", not for "linear-research-.." while you are for "log-research-..".
Actually there is still a very important "do i rush" question. . . any time you have a tech that's 'near' completion you have the choice of reducing focus on everything else temporarily to boost RP production to finish the Research sooner, so that you can start applying it's benefits sooner. by doing this you can see a net gain, over the course of several turns, but if you do it too soon, you end up stopping or slowing down everything else too much and fall behind on other area's of development. . . .

The Question is still there. It's always there. The 'Scale' of the 'cost' is slightly more subdued, and less overt, but it is still there.
M4lV wrote: In a system where RP are only generated by (even non-linearly) improvable research buildings, I don't have any problem, since it doesn't matter when I get more RPs, the tech projects are all finished the same way. Tech developments only mark waypoints on a clearly pre-determined curve.
Again not quite, if you produce 3 RP per researcher, then you finish something and produce 4 RP per researcher. Then you've got a VERY strongly dependent system (worse still if you pass from 2 -> 3 or 1->2). Making everything linear Doesn't get ride of this problem, adding a decaying exponetial term does not make things overly more complicated. (If the 'natural rate' produces 7 RP, and the Colony allotment is 32. . .. then the system is Linear. the 7RP just doesn't matter. )
M4lV wrote: It's the decision-making requirement that I'm complaining ;) about in your idea. I don't want to make any decision on each tech when to rush best, no matter if the required math behind it is multi-linear, logarithmic or non-standard analysis.
It's called strategy, it's a strategy game, controlling the known universe requires some thinking. I'd rather more then less (in the limit that excessive micromanaging is kept pointless). When a tech becomes rushable, you ask yourself when do i need this done by? (like you do with linear), If it's going to be done sooner you reduce or stop (Like linear) so that you can apply your resources to other projects (Like Linear). The only thing that is not like linear is that when you stop for a while your 'rewarded', and you can choose how long to stop, for how much 'award' you want. It's just another layer of depth, another layer of immersion.
M4lV wrote: Especially since getting the optimal point of RP investment is a joke to calculate and largely depends on "pi times my thumb", I don't want it in the game.
The optimal point is alway pi time your thumb, for linear, or exponetial decay. For both it depends (only) on how much you value finishing that research project vs apply the effort/money/people to other projects.

The 'optimum' point for both is 'Just before it's too late' which is in both cases incalculable, since 'too late' can mean just about anything, and just before means even less.
M4lV wrote: Btw. I do think my suggestions through, as I can even prove ;). Just have a look at the botf2 projects and mine/ours in particular and see how I think it is easiest and still interesting/variable enough. Don't want to make commercials for it but if you're curious to know what I mean in depth, go have a look at it :). I don't know much about Space Empires or SotS or SoaSE or GalCiv, but much about botf (and botf2). If you happen to have ST:BotF, you can also take the idea from how it's handled there though it was less sophisticated there..
Birth of the federation is botf, what is 'mine/ours'

I'm sure your capable of sitting down and thinking through to the application of your ideas. most people are. my apologies for having possibly suggested otherwise.

My point is that on a user level, for the rushing of a project the decay component does not change whether or not you have to think, only where your likely to put the balancing point. With every ship you build, every building you construct, every tech you research(linear or not) the same question always arises (doubly so when there is actually "Rush build" button), The Goal is to have as much in depth deep strategy as possible, with the min amount of micromanaging.


In the end the decision is up to the moderators and the people actually coding.

so that be that.
see you around the forums
Robbie Price.

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

Re: A Quick idea Re research;

#20 Post by M4lV »

as you might not have read, have a look at the thread here for what I meant with mine/ours: viewtopic.php?f=4&t=1794

I'm part of a team that creates a spiritual non-star-trek successor to botf and is pretty much through with everything except parts of the artwork and UI improvements which refrains us from announcing the game much around publicly since we're still not finished in replacing old botf structure pics (200x150 in size) that we use. Have a look at the BotE game for example. Supremacy can be played too but it has still 2 and a half years to catch up against us ;).

Anyway, of course there should be the possibility to effectively make rush decisions on (a small number of) techs but there shouldn't be as many penalties as there are in your system. With yours, I need to do the decision with each tech. With mine, I could just go and say, okay, I got 6 different tech areas, I uniformly distribute my RPs onto each of them and thus with the random rolls I do not get any penalties at all because 16% allocation to a tech means I got sufficiently enough turns left where my random rolls can always take effect effectively supporting such a uniform strategy (a uniform approach seems also realistic to me since you shouldn't be able to favor one particular tech field and neglect all others cause of the hidden interdependencies they have). When now my RP is increasing with upgrades, it doesn't matter since the increasement is divided by 6 in effects plus all surplus research percentages, i.e. when I'm at 95% research and produce 40% each turn, the remaining 35% are allocated automatically to the follow-up research project in the same tech area (or distributed uniformly to the others currently running). That resolves the rushing issue in my system.

So I can pretty much leave it at that and have no need to rush things hence can focus on other things in my empire and still have a pretty good research development. Of course those with the grips to manage it better can do that but rushing means less random-rolls-turns means penalties just as it should be. And that's very understandable to the user while the standard default "leave-me-alone-research-management" approach surely used by many players, inexperienced or not, is competable and does not punish those who use it too much.

Also, think about the AI. It has no thumb hence it needs to calculate our "joke calculation". When having so much input data like turn-wise-varying total RP production and so on, it is way easier to just let her uniformly develop her tech and make that approach a generally good and competing one like in my example leaving out the need to make rush decisions. Problem is hence if a standard rush decision point might be possible to calculate.

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;

#21 Post by eleazar »

Robbie.Price wrote:Actually there is still a very important "do i rush" question. . . any time you have a tech that's 'near' completion you have the choice of reducing focus on everything else temporarily to boost RP production to finish the Research sooner, so that you can start applying it's benefits sooner. by doing this you can see a net gain, over the course of several turns, but if you do it too soon, you end up stopping or slowing down everything else too much and fall behind on other area's of development. . . .
That will seldom if ever be a viable strategy. I wasn't around back then, but i think things were designed to eliminate rushing, and the need for continually tweaking the planetary foci.

Every tech has a maximum number of RP that can be put towards it per turn. You can't go faster than that. Generally you'll be able to research multiple techs at their max speed, and techs further down the queue will get the left-over RP.

Besides when you change focus, it takes time for that planet's production to ram up. I.E. when you change the primary focus of a planet from "mining" to "science", the mining drops off instantly, but it takes several turns for science production to pick up.

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

Re: A Quick idea Re research;

#22 Post by Robbie.Price »

Goodmorning all;

RE: "Rushing":

"Rushing" in this case is a particularly poor choice of words on my part. "Rushing" typically refers to the payment of large quanties of resources to acquire next turn something which normally would finish later. In this case the term was being used to distinguish between 'starting to invest RP As soon as a tech becomes available' Vs 'Waiting some number of turns before starting to invest RP'. I used the word rushing, because of the associated cost of starting early for the benefit of finishing early. I couldn't think of a better term, 'hurried, active, ... many others, not so much'


RE: RP/turn caps: I really appreciate the current system of RP caps, so that you can't get techs arbitrarily quickly, but if we were going to implement this system the Cap would have to be applied only to the applied Colony RP. Applying it to the full RP from both sources would result in Max every turn, and the immersion properties would all but be lost. (it wouldn't matter if you have 4% or 5% or a research treaty or not, if you virtually aways get max regardless[although the number of maxes you get would change so the dependency is still there, but weaker.])

Also if we were to implement this system the number of RP generated per turn would have to reduced to ~ 1/8th - 1/10th there current rates. The model treats research not as 'all the research that is going on on subject X' but 'the government funded specific portion of the research going on on subject X'. The 'natural' research will make up the vast bulk of the research, it's only the last part where the government steps in which colony RP would be applied, but that's just a balancing issue.

Re penalties in my system: I really don't see any penalties in my system, right now you have a list of tech you can research, you put them into a priority list, If you don't include a tech in the list yet, it doesn't accumulate RP, if it's not high enough, it doesn't get RP. If you keep that behavior then there IS no penalty since you add projects at the bottom, where they don't collect RP, and eventually they work their way to the top, by which point they start collecting RP,(if later then N, if you set N). IF you WANT a tech as soon as possible you put it at the top of the list and don't set N, BUT this does have a cost, in the end you pay more it's a trade off, strategic game play even.

It turns out that, as of right now, we don't uniformly distribute RP. So all the talk of distribution of RP becomes mute, unless we want to work out a system for distribution for FO.

Also the problem with 'follow up' technologies is that with non linear tech-trees such as FO on the Theory level(and other levels to lesser extents) there are no easily definable 'follow up tech'. Additionally with my system there REALLY isn't a follow up tech since you can't apply colony RP to the follow up tech until some period of time later.

Given the current distribution system, I assure can you that unless your RP production rate is insanely high you can safely ignore research.

Also I still think FO is against random rolls as a general policy, but you'd have to ask one of the coding team members. (+ rolling on a even distribution is never satisfying, to me.)

best wishes all

Robbie

Tsenzouken
Space Squid
Posts: 60
Joined: Sun May 25, 2008 10:15 am

Re: A Quick idea Re research;

#23 Post by Tsenzouken »

If I may interject... Computers are good at math. The average human isn't terribly. So, the mathematical complexity internal to the system is irrelevant (to some extent, as long as it doesn't slow the code down.) What matters is the apparent behavior that the model displays. If that behavior is coherent, logical and believable the goal is achieved and simplicity retained. Whichever internal system provides easier balancing without recourse to kludgy solutions is probably going to be best. =)

DISCLAIMER: I haven't done extensive mathematical modeling, so if you disagree with a number feel free to assist.

I personally favor a 'breakthrough' type system like what was used in MoO1. There is a certain minimum turns for a research, and the cost is gradually built up to. The difference I suggest is making it a non-binary system. To give you an idea:
RPc = RP cost of the technology
RPi = RPs invested into the research thus far.
X = RPi/RPc
Tech Scale3.PNG
Tech Scale3.PNG (13.14 KiB) Viewed 1303 times
Basically you end up with a tiny chance to get a breakthrough after ~40-50% completion which increases as you get closer to completion. You end up with a fairly decent chance to finish at the projected time, with the chance of a miraculous breakthrough or an overrun, both of which happen in reality and would be pleasant or unpleasant surprises, with a guaranteed breakthrough at 150% of the initial research cost.
The UI can display the "chance" in some way like MoO1 did, with 'lightbulbs' or the not-so-elegant MoO2 percentage. Numbers for visual representation and subject to change by people who are better at math than I.

I like the one on the left better personally. You can spend extra RPs to get the tech early, but you only really need to get it to complete and be patient with your scientists. Variables I would plan for:
Bonus/Penalty to completion ratio
Inefficiency (due to duplicated research, funding problems, etc)
The normal Research_Per_Pop = RPP * (1 + Bonuses / 100)
RPs generated at the empire level (some kind of Star Academy?) modified by empire tech level & gov't funding

EDIT: One intriguing thing which may in some part replicate the effect Robbie was going for: after 50% of the theory is researched, start letting players research applications. The benefit is that they can stuff more RPs into it early, the penalty is that their RP allocation into those applications is factored down by a percentage equal to 3/4 the completion of the project, defaulting to 1 after the theory tech is completely researched. It *allows* you to micro if that's what you want, but it would never be necessary to use it unless you were a more advanced player and willing to push extra hard to get something you wanted.

Thoughts?

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

Re: A Quick idea Re research;

#24 Post by M4lV »

yeah, that is the random rolls which exactly we are using ;).

As for the unlocking of application research before completion of necessary theory tech, I'd say I would give that approach a go but I would put a downside to it. If done so, chances are given that you get a malicious breakthrough in the opposite direction, i.e. all your research is lost because your experiments were based on a totally wrong theoretical assumption hence all application RP are lost. Basically the opposite thing towards the regular random rolls breakthrough with exactly opposite probability calculations.

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

Re: A Quick idea Re research;

#25 Post by Robbie.Price »

Goodmorning all,

I'm not completely against the idea of some that form of random roll, if distributed along an appropriate curve. At least the curve isn't linear, and by playing with the powers one could balance such a system reasonably easily.


But I'm curious:

Do you think that having a weak prerequisite system between techs is a good idea? If so how would you implement it?

How would you implement slow assistance/empire status as compared to other races, just RP?

Also i noticed that since your curves are strictly increasing that in your system Rushing a tech costs less... on average. The earlier i invest the more valuable the investment . . . since each turn after that has a slightly higher chance of random-finishing . . . .

If I were to use such a system I would have it start accumulating % chance to auto finish only after 90%, or even 95% of the RP's were invested, or as in moo2, after 100% was invested.

I would start with a rather high chance, say 20%, after 100% RP investment which then increases to 100% by (RP since user has been able to invest)*1.1 (with the exponetial decay system total RP's needed is quite high, compared to daily rate, 1.5 would be unreasonable.)

The other option would be to have any colony RP invested increase the chance that by forcing the research you put too many chefs in the kitchen and ruin the sufflé, and thereby loose some part of the accumulated RP. But then you've got three non-linear curves to consider... of one non-linear curve is too much for the user, three will really be over the top, n'est pas?

And; Is it worth while to have a system where there is a dynamically determined delay between completion of pre-requisites and being able to invest? I saw you mentioned it, but I didn't get how you intended to tie everything together to achieve fluidity?

IN short:

Are the goals I mentioned above worth while, and if so how does your proposal help achieve them?

I like the idea, as long as it's effects are limited to the last few % of the RP, and only vary the finishing time of a tech by only 2 or 3 turns on average, anything more then that and you really start to dig into the strategic focus of the game.

predetermination -> Sterile strategy (at least as sterile in the same sense as playing chess against a mediocre computer.)
general Predictability -> dynamic involving strategy
dominating random element -> strategic frustration or dumb luck (really really to be avoiding IMHO).

If a tech finishes 5 turns earlier or later then reasonably expected once in a game, that's a statistical fluctuation and it's involving. If i can never tell 5 turns in advance that I'm 80% likely to complete the tech before the 7th turn. . . that's a nightmare, or at least it would be for me.

Best wishes all,

Robbie Price

Tsenzouken
Space Squid
Posts: 60
Joined: Sun May 25, 2008 10:15 am

Re: A Quick idea Re research;

#26 Post by Tsenzouken »

Unlike industrial manufacturing, research is very rarely predictable in the sense of "we'll have <X> technology perfected within <Y> turns for <Z> RPs," at least at the theoretical level. Applications are somewhat more predictable, and refinements even more so. Remember that researching a theory has no immediate benefits (or few) so RPs spent researching a theory are largely an 'entry cost.' Researching applications while the theoretical basis had yet to be completed would be 'lossy' because of the factor value on the invested RPs. So 'rushing' a tech would actually be less advantageous. In this case it would be necessary for theory technologies to have a significantly higher RP cost relative to applications and refinements of the same average tech level.

I understand the concern about more always being better. My issue with 'free' resources is mostly that the system can provide significant benefits if you do the math involved. You might only have to touch it once, but you still need to do the initial calculation, and if it is indeed the most efficient means to get the research done most people won't ever change it, which leaves the system as being entirely deterministic just like we were trying to escape from to begin with.

More later, have to go pick up a tux. 8"}

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

Re: A Quick idea Re research;

#27 Post by Robbie.Price »

Goodmorning all,
Tsenzouken wrote:Unlike industrial manufacturing, research is very rarely predictable in the sense of "we'll have <X> technology perfected within <Y> turns for <Z> RPs," at least at the theoretical level. Applications are somewhat more predictable, and refinements even more so. Remember that researching a theory has no immediate benefits (or few) so RPs spent researching a theory are largely an 'entry cost.' Researching applications while the theoretical basis had yet to be completed would be 'lossy' because of the factor value on the invested RPs. So 'rushing' a tech would actually be less advantageous. In this case it would be necessary for theory technologies to have a significantly higher RP cost relative to applications and refinements of the same average tech level.
For the first part, i agree, for the most part technology advances are to some degree unpredictable, which is why i did and do like the idea of some(very limited) probability per turn, after accumulating enough RP. I would put that limit at 100, or very close to 100% of the techs RP. IMO, the strategic element of game play must remain dominant. FO is a Turned based strategy game, thus will probably end up having a much higher degree of predictability then real life, this is ok, since it makes the game fun. I would say that one should be able to 'guess'/'learn' to predict the turn of completion to within one turn no less then 67% of the time. (one sigma = 1 turn :- D)
Tsenzouken wrote: I understand the concern about more always being better. My issue with 'free' resources is mostly that the system can provide significant benefits if you do the math involved. You might only have to touch it once, but you still need to do the initial calculation, and if it is indeed the most efficient means to get the research done most people won't ever change it, which leaves the system as being entirely deterministic just like we were trying to escape from to begin with.

More later, have to go pick up a tux. 8"}
I'm not entirely sure what you mean here. . .

In my system as suggested the number of RP needed would be abnormally large, but the user would be blocked from directly contributing RP until the remaining cost dropped to a more 'classical' level. This, combined with a dynamically determined exponential rate which depends on a wide number of governmental/race choice factors, would determine how long the delay between finishing a techs pre-requisites and being able to start the next tech would be.

This could be replaced with a completely differnt system which involved no exponetial anything, if somebody wanted to really knuckle down and work one out. . . but the easiest method i can conceive of is the exponetial.

Once the tech is 'opened'/'unlocked' You have another dynamic, you can chose to research it the first turn it becomes 'unlocked' or wait. There wouldn't ever be one 'correct' answer, because some tech you really want sooner, others your content to wait for, sometimes it's of strategic importance, other times more pressing matters demand your attention. . .. it's is merely a tool to make it all more interesting and involving, at least as i see it. I acknowledge if not balanced correctly the amount of RP generated by the exponential can become quite large, but that, as i see it, is merely a balancing issue.

Perhaps I didn't understand your objection. . . but i hope I spoke to your concerns. . . Please elaborate.

From : viewtopic.php?f=6&t=2231&hilit=
Lupin III. wrote:
Finally there are the refinements. They could be the same thing as the applications itself, but you can keep researching them without end and no distinct levelups. For example you already have the application "Plasma Cannon", but you keep researching it and put RP into it. The more you research the better the stats (size goes down, damage up or whatever other stats there are). Of course the further advanced a refinement is the more RPs you need to get the same benefit (to reach 5% size decrease you need 10000RP, for 10% you need additional 20000RP). To make it clear, you wouldn't have a "5% smaller" technology to select, it's a continuous research.

These refinements should stack. So let's say "Plasma Weapons" is a prerequisite to "Plasma Cannon", refining "Plasma Weapons" would also increase the stats of the "Plasma Cannon". Some other refinements like "Miniaturization" would influence all parts or even planet improvements (Mini.) or "Compressed Plasma" (just making things up here ;-) ) would improve everything that has to do with plasma (Plasma Weapons, - Shields, - Reactors, aso.)
I would concur that the number of refinements should be, within limit, limitless. Which is to say, I would really like each additional weapon type to contribute something 'distinct' rather then simply be 'more powerful'. I would not however have a lack of distinct level-ups, because that makes it hard to judge where one tech ends and the next begins, and the rest of the techs are discrete, but it is a neat idea.

To follow up quickly on the 'Distinct' idea, ... For example if you had two weapons 'Lasers' and 'Disruptors' to give them arbitrary names. and let us say that 'Lasers' are the first weapons tech, and 'Disruptors' are the 5th. I would have lasers + 4 upgrades be able to cause slightly more raw damage/turn the un-upgraded 'Disruptors', but to compensate 'distruptors' would have an additional ability which lasers do not, while both would 'cost' about the same to produce. This would keep, to some degree, each of the research branches equally viable, in the end game. To some extent Moo did this, since future techs increased miniaturization of older techs, so you could pack more on a ship, and it was good.


I think your second point fits well with my concept of weak prerequisites: Where the 'natural' rate of development of a current tech you are working on depends on a weather or not you've finished any of a long list of ostensibly related technologies, refinements of previous techs would fit into this.


Best wishes both of you, and to all

Robbie Price

Tsenzouken
Space Squid
Posts: 60
Joined: Sun May 25, 2008 10:15 am

Re: A Quick idea Re research;

#28 Post by Tsenzouken »

Hrm. It seems like you believe that strategy = precision prediction. In my experience, that is not always necessarily the case. A scalar probability would still give you a general idea of when you would get a theory breakthrough, but not pin it down. I do agree that the numbers probably need to be changed, but there is a specific reason that I allowed for significant overruns in theoretical research.
If theoretical research takes a long time, it prolongs the "ages" of technology. A research-focused empire will probably get to the applications and refinements faster, but shouldn't be able to just run through all of the theoretical techs and leap from Laser Mk1 to Quantum Phase Nanocannon Mk60 in one jump. That is why in the diagram the guaranteed-discovery point is at the 150% RPs invested to cost ratio.
Essentially, if you know you'll get it eventually through random chance if you keep steady progress, you'll be more likely to start researching applications and refinements as opposed to lumping all your RPs into that one theory to finish it and move on to the next. With any luck, making it disadvantageous (in general) to do that (that is, lump all RPs into one project) and better to refine existing technology for a time will make it easier to balance tech.

The exponential percentage method seems like a more complicated way of requiring a minimum turn length on research. Granted, probably a more elegant way of requiring minimum turns--but still such a method. Though, come to think of it, the numbers could be manipulated such that your chance to make a breakthrough began at 66.67% of the way through the research, but you could start researching applications as soon as you got 33.3% into the theory, using the curve from my earlier graph as a basis for RP factoring. The end result would be a theory which takes a long time to complete, and a use for raw RP output to be channeled into other things at the same time.

I also like the idea of an inefficiency factor built into researching, which things like Semi-sentient Databases would correct as Learning research progressed. That way not ALL research techs have to be +X max research infrastructure meter. Just a thought.

P.S. I also like 'weak' pre-reqs. Trying to figure out an easy way to implement them, more on that later.

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

Re: A Quick idea Re research;

#29 Post by Robbie.Price »

Tsenzouken wrote:Hrm. It seems like you believe that strategy = precision prediction.
I would not go as far to say Precision, but one of the key quotes for war "Know thy Enemy" assumes that the knowledge is meaningful, on account of the behavior being to some degree predictable.

I would not have 100% predictability, since i find that rather dry, but rather 'predictable to within the scope long range plans' is important. Even big, game changing, upsets are cool, from time to time, since they force the user to rethink everything. But having something as ubiquitous as research finishing times be overly jumpy would only cause me stress and headache. (key being overly, some is nice of course)
Tsenzouken wrote: In my experience, that is not always necessarily the case. A scalar probability would still give you a general idea of when you would get a theory breakthrough, but not pin it down. I do agree that the numbers probably need to be changed, but there is a specific reason that I allowed for significant overruns in theoretical research.
If theoretical research takes a long time, it prolongs the "ages" of technology. A research-focused empire will probably get to the applications and refinements faster, but shouldn't be able to just run through all of the theoretical techs and leap from Laser Mk1 to Quantum Phase Nanocannon Mk60 in one jump. That is why in the diagram the guaranteed-discovery point is at the 150% RPs invested to cost ratio.
I definitely concur jumping from lasers 1 to anything mark 60 is ludicrous, but that is far, far removed from anything i was even remotely suggesting.

I would restate for the record that if using an exponential system the 150% marker would be unreasonably high, since the user can only direct focus for the final %ages of the total research costs anyway, the vast majority being 'filler' to give the idea of the correct time between techs. Unless you defined it as 150% of the amount 'since the user was able to direct RP'. But this is a mere technical detail.
Tsenzouken wrote: The exponential percentage method seems like a more complicated way of requiring a minimum turn length on research. Granted, probably a more elegant way of requiring minimum turns--but still such a method. Though, come to think of it, the numbers could be manipulated such that your chance to make a breakthrough began at 66.67% of the way through the research, but you could start researching applications as soon as you got 33.3% into the theory, using the curve from my earlier graph as a basis for RP factoring. The end result would be a theory which takes a long time to complete, and a use for raw RP output to be channeled into other things at the same time.
I would very much be interested in a simpler method for requiring min turns before applying RP, which allows all the dynamic behaviors which i have outlined as being possible using a simple exponential decay system. I have yet to find one.

But that aside, I would have the, can direct RP, level around the 80% completion point, (refinement techs earlier theory techs later). I would then have the probability to self complete start no earlier then 95%, or even 100%.

Having the threshold for direction so close to finishing permits the 'between techs' period to be as flexible as possible, since it has very few other game effects; it simply allows the maximal room for balancing, for the minimal coding/intellectual effort, and minimizes some of the more undesirable, by some, effects of non-liner systems.

Also the question of 'where' the probability of completion starts to be non-zero is somewhat academic, considering the 'zero' point and the '100%' points are both arbitrary. Somewhere between investing no RP and investing Lots of RP, the probability of finishing a tech next turn goes from pure 0 to some non zero. If you define the 'complete' point as being a 'complete enough' point then it just makes the most intrinsic sense to the user that after reaching 'complete enough' you have a chance of finishing the tech, before that you don't. Crossing the 100% finished line aught have some distinct effect, at least in my mind, and I'm sure I'm not alone. That can either be 'finishing' the tech, or now having the possibility to finish the tech. But it aught be somewhat important either way. (otherwise why does it exist?)
Tsenzouken wrote: I also like the idea of an inefficiency factor built into researching, which things like Semi-sentient Databases would correct as Learning research progressed. That way not ALL research techs have to be +X max research infrastructure meter. Just a thought.
I'm afraid I do not understand this point. Although the most common effect may be +X max meter, +X% to natural research rate, would also be common, at least with an exponetial system.

could you clarify for me, thank you.
Tsenzouken wrote: P.S. I also like 'weak' pre-reqs. Trying to figure out an easy way to implement them, more on that later.
ok

Best wishes, thank you all,

Robbie Price.

Post Reply