Public Review: Tech Tree I

[Tyreth]

It's been a while since I've been in the tech discussion, and I see a lot has happened. Been rather busy over January period.

Categories should be as directly related to gameplay and particularly strategy as possible (i.e. no 'mathematics' category, but 'Ships' might be o.k.) (open question)

My preference is for more general names like Mathematics, Chemistry, etc. Two reasons:
1. It allows us to come up with creative advancement names like "Super Tensile Solids" rather than "Super-Surreal Laser". That way we can give the genuine feel that the player is advancing through theoretical research in a real universe, discovering its secrets. I really appreciated the effort of names and quotes that went into the SMAC tree.
2. Making theoretical advancements really about the underlying scientific advancement, rather than the direct application, lends itself naturally to a more intertwined tech tree. Some may not like that, I prefer it.

Research cost will scale with galaxy size. (open question)

Make this an option. When players select different galaxy sizes, a slider automatically adjusts itself to the recommended tech speed. The player can then slide it to a default (mid) speed, or wherever they like.

Randomly block out different techs each game or not?

Definitely not. The reasons Tzlaine mentioned are good for me. However, I have a real distaste for randomness like this. People love to discover patterns, imo. One tech should necessarily lead to another.

Will it be possible to develop applications without knowing the theory, if all the prerequisites have been met, maybe at some kind of penalty?

No.

Do we want to repeat the outline of the tech tree to extend the research game?

Unnecessary if we make the tree sufficiently hard to complete in its entirety. SMAC in my opinion was made in such a way that the game couldn't help but draw near the end once the tech tree was nearly completed. By this time at least one faction had it in their hands to win. We should balance the game in this way, so that it feels like a story unfolding rather than a repeat of the same numbers.

Do we want to achieve multiple paths by balancing all techs perfectly, or by deliberately introducing several 'prefered paths' or tech families? AND How can we produce genuinely alternative gameplay styles and support this with the tech system? AND How do we do this whilst keeping things relatively simple and fun?

We should have several preferred paths. One way to balance this is to have a triangle, with each point being a defeater of the next, and being defeated by the previous. This means a player has a benefit in focussing (unless he's a strong tech race that does not need to focus) in one of three preferred paths.
This can also help add to the theme of the game. Example of three preferred paths:
Path a, Harmonizer - this tech path promotes technology that works in harmony with the universe, galaxies, systems, planets and even down to the atomic level. These techs presere rather than destroy life and the ecosystem(?) of the galaxy. Example technologies would be strong terraforming abilities towards greener lifestyles. Biological ships rather than pure technology ones.
Path b, Pillager - this tech promotes a lifestyle of plundering and taking what is needed for progress. As a result, this technology tends to destroy more. Strong destructive spy techs, ships that tend to mutilate the life on a planet, and so on.
Path c, Diplomat - this path is a lifestyle of defence, and patient necessity. The player may find himself with good social techs, nice strong weapons, an excellent infrastructure on his planets.

Anyway, such paths could be fleshed out, and their exact boundaries will not be fixed, but rather that players may find themselves going along one of these preferred paths to get the techs that build on each other. In SMAC (yes, again!) you could terraform your territory with industrial improvements, with trees, or with xenofungus. Each had its own limitations and advantages. The player who was playing a green tech research tended towards natural armies of mind worms, locusts, etc - and these were very different in war. So there was more than one preferred path, but enough variety that each nation was different.

On the topic of the nature of the tech tree, one element I think important to keep in mind is the varying stages of the game. In the Civ series, there were distinct markers up the tech tree that resulted in a radically different game. In SMAC the same thing happened, but the markers were not so distinct. The discovery of certain technologies by one race necessitated the discovery of it and its counters by other races. Such was good to give the game a feel of distinct progress. This was discussed on the old forums, but I cannot recall the details.
I like this, and its one of the main reasons I don't like the idea of a recursive tech tree.

We need to think what to do about the spare RP's left over each turn (either another use for them (is this the place for refinement?) or bear in mind that they will need to be easily redirected back into the economy somehow.)

I've thought of two solutions:
1. In the research screen you have an option of where to send the excess RP for conversion - minerals or industry. All excess rp's are converted over to that resource at a fixed rate (eg, 0.5). If we want an extra racial advantage or an option for research to modify, then the RP conversion rate could be variable. eg, one race gets a 0.75 multiplier for conversion to industry. This just represents excess research going into pet projects of scientists with resources the government grants but can't direct.
2. Reverse the above - no such thing as directly produced RP's. You start however many projects you want, and the RP's are drawn from existing industry or minerals (either by choice or we make it fixed). The conversion rate is set by racial stats and by planets, and will be fractional (eg 1.26). This means you will only ever spend exactly what you want, and you can start as many projects as you want, crippling your industry as much as you are willing.
The second solution is different from the norm in these sorts of games, and I personally don't like such precise fractional conversion rates. Therefore, I favour my first suggestion.

[Bastian-Bux]

Well, I go with Tyreth on that .

[Aquitaine]

May I ask what the point of cancelling projects is then? (other than uncluttering the UI ).

I don't think I've ever canceled a project in HoI. And I can't really imagine why you would, since you can suspend it and keep all your progress.

[Aquitaine]

1. It allows us to come up with creative advancement names like "Super Tensile Solids" rather than "Super-Surreal Laser". That way we can give the genuine feel that the player is advancing through theoretical research in a real universe, discovering its secrets. I really appreciated the effort of names and quotes that went into the SMAC tree.

I actually agree with you about names but disagree on categories. There is no reason that we cannot have super-tensile solids and other things like that with more specific categories. Typically these would be the names of theories rather than applications, but they can be both; HoI does it, certainly.

Aqui "many categories" taine

[PowerCrazy]

haha. Good ol' SMAC. I agree with tyreth about the diversity and generality of Tech categories and individual tech names. But as far as extra RP. I say we have a third branch "refinement" that would work exactly like tech worked in moo1, moo2 etc. I would click on the refinement window of my tech place and then select "lasers" or "nuclear missiles" whatever. Than the more excess RPs I have a turn the more advanced I get with my nuclear missiles. So after 3-4 "breakthroughs" in refinement I would be able to have (using moo2 terminology) fast, eccm, mirv nuclear missiles.

Then we could also give bonuses to future refinement depending on how much you refined the lvls below. Thus it would always be beneficial to refine techs to a certain lvl. And at end-game when i'm a massive conglomerated empire I'll never run out of things to research. I can refine my ship engines, my shields, my beam weapons, my PD weapons, my missiles, etc. Almost unlimited.

And to keep weapons from getting "too powerful" we can have diminshing returns, such that you get the most benefit from the first 2-3 lvls of refinement. But if you have the resources you can still get a small advantage by continually refining a weapon, at least until the next weapon project finishes. Then you will want to start refining the next one.

[Sandlapper]

Tyreth wrote:

Quote:

Randomly block out different techs each game or not?


Definitely not. The reasons Tzlaine mentioned are good for me. However, I have a real distaste for randomness like this. People love to discover patterns, imo. One tech should necessarily lead to another.

People love to discover patterns, imo. This is exactly, the arguement for randomness. A set path will be discovered, then what, you play this over and over again. That why I prefer randomness, not knowing what the path will be before I even start the game.

Being that their seems to be a consensus for a static tree; Why not allow a preselection of tech through the tech tree from beginning to end, at the start of the game? Preselect into a que, wait for RPs to build, wait for selections to run their allotted time, then provide a popup when they are ready to be utilized. An advanced player could then concentrate on strategy and ignore the tech tree.

PowerCrazy wrote:

But as far as extra RP. I say we have a third branch "refinement" that would work exactly like tech worked in moo1, moo2 etc. I would click on the refinement window of my tech place and then select "lasers" or "nuclear missiles" whatever. Than the more excess RPs I have a turn the more advanced I get with my nuclear missiles. So after 3-4 "breakthroughs" in refinement I would be able to have (using moo2 terminology) fast, eccm, mirv nuclear missiles.

I concur with a refinement approach such as this.

[emrys]

Tyreth wrote:
Being that their seems to be a consensus for a static tree; Why not allow a preselection of tech through the tech tree from beginning to end, at the start of the game? Preselect into a que, wait for RPs to build, wait for selections to run their allotted time, then provide a popup when they are ready to be utilized. An advanced player could then concentrate on strategy and ignore the tech tree.

I do see the point you're driving at here, and if did becomet he way people played the game, it would be a very sad result. That said, I think the best way of getting people to think about their tech choices anew each time they play is by making sure that the [/b]game situation is the strongest determiner of what are the best choices to make at a given point. If that isn't the case then a randomly holed tech tree is at best only going to make players think "damn no uber laser this time, oh well, let's make do with these other things for a few turns until I can get to the next step in my standard plan, and that'd better be there or I'm scrapping this game."

[emrys]

And to keep weapons from getting "too powerful" we can have diminshing returns, such that you get the most benefit from the first 2-3 lvls of refinement. But if you have the resources you can still get a small advantage by continually refining a weapon, at least until the next weapon project finishes. Then you will want to start refining the next one.

As a side point, I quite like the idea of having a heavily refined older weapon being slightly better value (but different qualities) than a 'bleeding edge' brand spanking new unrefined one. So there would be a small dilema as to whether to stick with the tried and trusted weapon, or start off down the new path (which would rapidly become better value).

[tzlaine]

I see that some here have said they want two coexisting tech systems, one of which is like HOI (for theories and applications), and one that is like MOO1/2 (for refinements, possibly only using excess RPs left over from the HOI-style spending).

I think this is completely wrong. That would mean building two UIs, two sets of logic in the code, and having the user research new techs in two ways. This obviously violates KISS. We should figure out how we want to conduct research, and do all of it in a uniform manner.

I just stated why I want to do things this way, but I have no idea why some feel that two systems should be used. What is the motivation for doing things the other way?

[emrys]

I see that some here have said they want two coexisting tech systems, one of which is like HOI (for theories and applications), and one that is like MOO1/2 (for refinements, possibly only using excess RPs left over from the HOI-style spending).

I think this is completely wrong. That would mean building two UIs, two sets of logic in the code, and having the user research new techs in two ways. This obviously violates KISS. We should figure out how we want to conduct research, and do all of it in a uniform manner.

I just stated why I want to do things this way, but I have no idea why some feel that two systems should be used. What is the motivation for doing things the other way?

I at least seem them as co-existing within the same screen of the UI, with refinement being launched from the inevitably needed list of researched techs. Refinement in this way will however admittedly need some additional UI and a fair bit of extra logic code though.

The user would NOT research new techs in two ways, because refinements would NOT be new techs. They are specifically enhancements to existing techs.

The motivation for doing things this way:

Assumptions:

1) FO will have a similar form of underlying economy to the MOO series.
Particularly there will be RP produced some way, in a fashion that will generally increase with time and empire expansion and will be measured in numbers above the tens (probably orders of magnitude above).

2) The RP production of an empire may however drop suddenly (e.g. if a major research world is conquered or destroyed).

3) We are going to be using a HOI style tech research model (by which I mean research will requiere consistent expenditure of X RP for Y turns to research a project). X is a number which is NOT 1. This model has a variety of useful benefits, including limiting tech runaway, encouraging planning and forethough in tech choices etc. and being aesthetically pleasing (see below).

4) For some reason it will not a trivial task to consistently produce a precise level of RP without a waste of some kind (productive potential, efficiency or turn advantage). [In practice this is likely in any model that encourages investment in a particular strategy or discourages micromanagement]

Problems Forseen:

1) Quantisation loss
Take a five turn snapshot. An Empire produces the following RP on each turn,
100, 105, 109, 112 and then 116

There are projects available to research at the following costs
A=50, B=30, C=20 and D=10, each takes more than five turns.

so empire runs these projects:

Code: Select all

RP in  projects  cost  surplus  cumulative surplus
       running
100    ABC      100    0               0
105    ABC      100    5               5
109    ABC      100    9              14
112    ABCD     110    2              16
116    ABCD     110    6              22

If the 'surplus' is not stockpiled (i.e. basic HOI model) and is wasted this is a quantisation loss which occurs because projects expenditure only comes in discrete chunks, and the chunks are larger than the production chunks.

This is a worrying temptation to micromanagement.

2) over capacity loss:

If the tech tree doesn't unfold quickly enough, there may be situations where a player has an RP income which he just cannot spend. If in the above situation project A had finished after two turns, but had only opened up a single 10RP for six turns project (E), you would have:

Code: Select all

RP in  projects  cost  surplus  cumulative surplus
       running
100    ABC      100    0               0
105    ABC      100    5               5
109    BCDE     70    39              44
112    BCDE     70    42              86
116    BCDE     70    46             132

Most of the loss here is from the fact that the player simply cannot spend the RP at the rate they are capable of producing them. i.e. an overcapacity loss.

This could be dealt with by saying "tough, cut down on capacity then". Note however that if suddenly a whole bunch of projects or some very expensive ones open up, they'd need that capacity you've just demanded the player throw away.

N.B. I'm sure there was a third problem I had in mind when I started writing this, but I can't think of it now, so expect an edit when it comes to mind.


Solutions:

There were two ideas that leapt to mind

A) Ensure there are never wasted RP, by allowing them to be redirected somewhere else at no loss(to avoid waste) and automatically(to avoid micromanagement). It is hard to see this working with any system that encourages investment in Research production (as opposed to something else, e.g. ship building or colonisation), in other words you'd probably have to have RP produced simply by spending (money/production/chocolate) at an empire wide level with no research buildings, infrastructure or specialisation of planets. I.e. one big "research slider".

OR

B) Give you something else to do with the left over RP to smooth out the jumps in expenditure. This requires something you can spend a flexible amount of RP on (in as small as 1 unit), and get some return for (though not necessarily immediately) There were two ideas here:

Either

Bi) the spare RP builds up in a stockpile until you can buy one of the normal projects "upfront" i.e you have X*Y RP. You then spend that RP and wait the Y turns. This might solve quantisation loss, but would probably be less effective at solving overcapacity loss, since you'd eventually get to the point where you were just buying everything up front. In addtion part of the benefit of the HOI model is that you have to plan future expenditure since you have to maintain the funding of projects. Also there is the aesthetic appeal of a system where you have to chose a project and research that, not just 'do research' and then choose what is was you were researching. That would be lost with this change.

Or

Bii) the spare RP is spent as it is produced on something, that something acts like either a continuous fractional improvement or a MOO3 style research process with results when you hit certain levels of RP invested. The natural choice for this would be improvements to the stats of existing applications, what we've been calling "Refinement" but could equally be called the "Development" of R&D or "Engineering" as opposed to "Science Research".

This seems to be complementary to the HOI model (HOI model:Pure science theory takes a certain time and research effort to understand because people have to talk, discuss and absorb stuff as they work on it. Once theory is reasonably understood, ideas for applied science Applications begin to seem plausible rather than pipe dreams, but they take the same process to bring to fruition. Extension: Once you have a working prototype sitting on your desk, improving it becomes more a function of the amount of people working, and money spent, on improving it, and less of the time taken to absorb and promulgate the ideas, so you can do Development just by piling on the RP.)

Given that Bii) is complementary to the HOI model, but seems to extend it to solve the problems that might probably come up in FO, without putting many constraints on the rest of the design, I personally think it seems like a good idea.

On the other hand if we can come up with a simpler system that keeps the benefits of the HOI model whilst avoiding the difficulties I can see are likely, and without adding any of it's own, then I'll probably be all for it.

[DemoMonkey]

Interesting idea. A further option, Biii) if you will.

"Spare" RP go into a separate Tech Field that cannot be researched normally called "Consumer Tech". Each level obtained provides a small but welcome empire wide morale/production boost.

Basically, in the course of developing Space Travel, Uberweapons, and Things Man Was Not Meant To Know, your scientists also happen to stumble across teflon frying pans, Tang, and laser eye surgery.

No wasted points and no micromanagement.

[PowerCrazy]

hmm. Biii seems interesting, BUT it takes away the options that refinement offers. By allowign a player to take old weapons and make them cheaper, take up less space, and be marginally more effective it gives more options for ship construction. Also certain weapons would be a natural choice to refine.

A Moo2 example. Your fighters used your best point defense beam. The best point defense was particle beam, and the first was laser beam. Obviously if you got a better PD beam your fighters would be more effective. So using the refinement approach. You are in mid-game and you have carrier fleets. Well the enemy just upgraded their armor so your carrier fleets are less effective now. So what do you do? You could scrap the fleets, change your tactics, etc. Or you could refine your latest beam weapon and make it point defense capable, OR you oculd refine the current PD beam and make it smaller/cheaper and allow your fighters to carry twice as many, or have twice as many fighters/carrier. Etc.

In this example there is no clear better way but all those options are there.

My main problem with Biii is that the benefit gained by extra research is intangible. Its an unknown, and some players might find that disconcerting. I want the reasearch I do to have well-defined results. Also there wouldn't be Micromanaging with the refinement option, becasue you simply click on which areas you want to refine and viola wait a few turns and there ya go, one refinement breakthrough. Then if you odn't change anything it will continue to refine that area. And of course we oculd allow players to have focused refinement... Instead of refining a field you could refine a certain weapon only, or a missile or a ship engine, etc. The possibilites are endless.

[utilae]

I would rather excess rp go into something like Biii then refinements. Because you have no control over your excess rp, so you would have no control over your refinements. But Biii. It would allow secret 'bonus techs' to become available. So the excess rp could be used well. You would not have to concentrate on gaining these bonus techs through excess rp, because it is the uncertainty of getting a bonus tech through excess rp that will make it fun. In a sense it would give a slight randomness to a static tech tree.

As for refinments, well we obviouslt have to keep them, so we can have refinements side by side with this Biii idea. Though I would really like to see a slider where you can distribute rp between research (applied and theory, excess go into random bonus tech) and refinements.

So once I have a few basic techs, ie lasers, then I can forget getting better weapons like plasma cannon. I can set all my rp to go into refinements, so that my lasers can be upgraded heavily.

I think that a tech (eg laser) should have a level, so level 0 is theory (ie not applied yet), level 1 is applied (ie it is now applied) and level 2+ is refined (ie applied but refined as well). So all you would be doing in the refinement screen improving an applied tech by 1 level, each time the tech is refined (leave on auto it will continue refining).

Maybe as well certain techs refine themselves in different ways. If a tech goes up a level we could say that certain stats improve:
eg
Damage
Size
Cost
Range
Armour (option becomes available at level X)
Armour Penetration (option becomes available at level X)

When a laser is refined, its damage may not improve as much as a plasma cannon being refined, but the laser may become smaller and cheaper easier then a plasma cannon being refined.

Also certain optional refinement extras such as an armoured missile or armour or shields penetration, would become available at a certain level. So a laser may reach level 20, then armour penetration can be used as an option with your refined laser, when you design sihps.

[Aquitaine]

Because you have no control over your excess rp, so you would have no control over your refinements.

This is incorrect. Your excess RP is whatever you don't spend on projects. So if you want more excess RP, you don't start new projects or you suspend existing ones.

[tzlaine]

This is in response to emry's post. I just didn't want to requote the whole thing.

I think you are possibly asking the wrong question. The principle question you seem to be addressing is "How do we as game designers ensure that the RPs a player produces in any game get fully used?". I'm not sure why we even need to ask this question, much less answer it. It is part of any 4X game that the player must develop the areas where he can realize the most gain. For instance, in the early game you might want to ignore tech for production, then ramp up tech later. In the midgame you might want to put all your efforts towards getting a key tech, and so basically stop all your production in an effort to do so. These kinds of choices are what make the game.

If a player decides to go full-tilt after tech, but neglects to research a broad enough range of techs to leave enough simultaneous projects in the future, he needs to get better at the game, we don't need to redesign the game.

By having refinements as just smaller projects that each affect a single application, you will almost always find something to take up those last few RPs you have left over from the big theory and application projects.

It might remove the micromanagement aspect of this model to be able to define an arbitrarily-long queue of tech projects. We could even make the queue smart enough to look past the front of the line if one of the later projects could be started now. That way, "excess" RPs get spent automatically, but we don't have parallel research models. I still haven't heard any compelling reason for having two models.