Tech Research Chart 2

[Mark_Everson]

I think its a Great idea too... I think with the basic tech of Religion, you might be able to get by with only 2 ancillary techs, polytheism, and Monotheism. These would be prerequisites for the basic tech. When you have them your religion value is easier to increase. Philosophy might be another prerequisite.

Sounds Great,

Mark

[Lord God Jinnai]

Maybe not that few...i'd like to include some eastern types which wouldn't conform to either of those. Buddhism is one which i'm using Way (or path) of enlightenment and taoism is the other which requires several, but these also tie into tai-chi...those are harmony with self and nature (2 diff) and non-inteferance and theory of change...i might drop 1 though.

[Mark_Everson]

Hi LGJ:

I think some of the tech model ideas are a next logical thing to try to get feedback on in the demos.

I can (or maybe Anton, who expressed interest) fairly quickly modify my old tech model and interface for the dry run. Once we're satisfied that the new tech model works ok, then we can do a full-blown model and start work on an interface.

The things we need to figure out before the demo 4 version of tech are what the cost profile will be for basic techs, and of course specific costs and prerequisites. For the basic tech cost profile I think something that costs exponentially more the higher the level should work well. That's because civ growth will also be exponential.

Probably we could use basic techs in the army area, production, and maybe something like metallurgy. Further suggestions?

Specific techs could be flexible units, iron working (together allowing Legion-type flexible heavy infantry) and maybe a few others.

What do you think?

[Mark_Everson]

LGJ and other interested parties:

I'm going to post two things here today relating to the Tech model. The first one, in this post, is a proposal for the specific way we should handle basic techs. Since Anton/rentgen is interested in coding up the model we need to get some specifics going for how things work. I thought about it some yesterday, and I'll present what I came up with.

The second post will be a proposal for how to use the approach in this post to take a first shot at coding up a small tech model for demo 3. That way we can hopefully get some feedback about how people like the general aspects of the Tech model, which I think is Really Good :-)

Proposal for how to handle basic techs...

Assumptions:

Personally, I think only the 60 or so most crucial achievements of humanity should be regular techs. And All the regular techs should allow you to Build or Do something Really new. Basic techs should also have some techs embedded in them IMO. This would allow for some advances that are pretty much on the standard track to be covered, without them being a whole separate entry. This would also add some description to the otherwise bland basic techs.
I am taking LGJ's approach to basic techs in that they will simply have a percentage rating. A basic tech at the 0% level is essentially neolithic technology. 100% is the modern level. The apogee of ancient civilizations might be something like 20%. Technology past modern level just goes past 100%. Regular techs can have prerequisites such as Metallurgy 5%. Units, infrastructure, etc. that is built in the game will generally have one, perhaps two, basic techs associated with it. The unit description would contain a level of the basic tech that affects its cost for deficiency as outlined by LGJ above. The usual span for bonuses or penalties to unit cost or capabilities would be something like plus or minus five percent from the basic level required. So suppose I'm building a Legion which has the associated basic tech Metallurgy 10%. If my metallurgy were 15% instead I might get a 50% bonus off the cost of a legion. Similarly, if my metallurgy were 5%, my legion would be almost useless (They are flexible and well-trained, but with weapons that break every time they use them).

My main thrust:

I think the way prerequisites work as discussed in the thread above is pretty good. What I'm primarily trying to address here is how expensive the different percentage levels of the basic techs should be. If we have a good handle on these point costs, then it's also easier to figure out the point costs for regular techs at the same each level of technology. Seeing as I'm a physics nerd, my approach to this is to use scaling arguments.
I'm assuming for playability reasons that we would like to have the technological progress somewhat even throughout the game. Clearly it can't be too fast in ancient times, but there's a lot of early ground-breaking stuff for the player to get then so I'm not too worried. Taking as a premise that we want things to be fairly even throughout history, what can we say? I need one more assumption, which is that technological progress, tech points if you will, scales roughly as population times per capita income. That is, GDP. In the range from subsistence to a developed modern economy per capita income changes by about a factor of 50. Population also grows by more than a factor of 10 going from ancient empires to large modern states. To give a round number let's call it a factor of 20. So GDP should grow by about a factor of 1000 over the game. This large factor makes an exponential or power law most suitable. Before bringing the prerequisites for basic techs in we can produce the desired behavior by having to cost increase by a factor of two every 10 percent. Ten doublings gives us to factor of 1000.

Temporary conclusions (2 will be changed later)
1) tech points earned per turn grow by about a factor of 1000 over the game
2) for every 10% increase in basic tech there is a doubling in the cost

This would work fairly well, except the way we are planning to use prerequisites for basic techs distorts the system. That's because each point in the basic techs early in the game costs a lot because you don't have many of the prerequisites. I reproduce my post above here since it is critical to what goes on later. It is talking about application to a regular tech, but using prerequisites for a basic tech would work similarly.

Instead of having hard and soft prerequisites for techs, why not just have a weighting put on all tech prereqs. Then we can tune them to be as hard or soft as we want. The weighting factor would affect the cost to put one point into getting the tech. For purposes of demonstration just assume each prereq has a weight which divides the points put into the tech if you don't have that prereq, and that all the modifiers are multiplied together.

Suppose we have the tech Navigation with the pre-requisites Astronomy (formerly hard), Compass (formerly soft) and Seafaring (formerly soft). In my proposed method it might look like this (don't worry about the specific numbers, this is just for demo):

Navigation (100points)
Astronomy (5)
Compass (2)
Seafaring(3)

So if you have none of the pre-reqs each point of Navigation would cost 5x2x3 = 30points. So all of Navigation would cost the Enormous amount of 3000 idea points. If you have Astronomy and Seafaring it would cost 200. Of course with all three pre-reqs the cost would be the base cost of 100.

Just shooting from the hip, it seems each basic tech will probably have about 10 prerequisites spread across history, and each will impose a penalty of about of factor of two if you don't have it. So, earning your very first percent in a basic tech you are stuck with about 1000x in penalties for each point. For instance in Engineering you don't yet have the wheel, simple machines, mechanisms, the steam engine, and a bunch of other stuff. So, we need to make the costs for the low end of basic techs cheaper to compensate for all this baggage. If my assumption about 10 prerequisites, each with about a factor of two penalty is correct, then we just need to modify conclusion 2 above just slightly. If we put in that doubling every 5% instead of 10% it will compensate for the prerequisite effect. See the examples below.

Conclusions
1) tech points earned per turn grow by about a factor of 1000 over the game
2) for every 5% increase in basic tech there is a doubling in the cost
(this may need to change as we determine how many prerequisites each basic tech has, and what it's penalty is)

Example (don't take the specifics seriously, I just put them down quickly):

Basic Tech Engineering
'Standard' cost: 0.01 idea point (this needs to be a small number, you'll see later why)
prerequisites:
the wheel (2x)
simple machines (2x)
mechanisms (2x)
education level something... (not a tech, determined by education of people) (2x)
invention (2x)
steam engine (2x)
Metallurgy 40% (2x) (can't make much progress beyond a certain point without decent metallurgy)
innovation (2x)
research and development (2x)
miniaturization (2x)

Given this as a basis, here's some examples.
Starting with the neolithic Engineering level of 0% the cost to go to 1% is about 10 points (standard cost times about 1000 = 2^10 because you have none of the prerequisites and each one imposes a penalty of a factor of two).

If you are at 19% and have the first 2 prerequisites what is the cost? First, if you had none of the prerequisites it would be the standard cost (0.01) times 1000 (no prerequisites yet) times 16 (a factor of two for each 5% in the basic tech = 2^ 4). This gives a total cost of 160. If you have each of two prerequisite techs this will lower the cost by a factor of two for each which gives 40.

How about obtaining 50% from 49%. First, what if you had no prerequisites? (This is quite a stretch, since if you had no prerequisites you would probably never have gotten this far) the cost would be about 10,000. With the first three prerequisites the cost would be about 1200. With the first five prerequisites (or any five for that matter) it would be about 300. If you had just come into contact with a benevolent advanced civilization who had given you eight of the prerequisites it would only cost you 40!

Getting from 99% to 100% with all 10 prerequisites already achieved would cost about 10,000 idea points. If you were lacking 2 prerequisites the cost would be 40,000

To summarize:
%, Prereqs Obtained, Cost
1%, 0, 10
20%, 0, 160
20%, 2, 40
50%, 0, 10,000 (just about impossible at that point)
50%, 3, 1200
50%, 5, 300
50%, 8, 40
100%, 8, 40,000
100%, 10, 10,000

Of course the standard cost numbers will need to be balanced based upon how hard an idea point is to generate, and the value of that basic technology to the player. Based upon this limited example, I think my assumption of 10 prerequisites is probably too small. However, the basic ideas are still relevant.

LGJ, and others, what do you think?

[Keli]

Mark, you have a real knack to describe farily complex things in 'layman terms' .

Anyway, I think your assumptions above are good. I haven't studied this thread thoroughly, but I was wondering: Why do the prereqs for a given base tech have to apply from the start? Maybe this question is answered somewhere in the thread, but here it is anyway. I can think of several ways to handle this thing differently: a) if each of the 10 prereqs (or whatever the total number is) is dependant on each other, then only the 'topmost' (newest) is used as prereq, this can maybe be linked to a specific base tech percentage; b) each prereq only modifies a specific range (like 10-20%) of the base tech percentage; c) each prereq has a ceiling to how high the base tech can rise.

OK, none of the above ideas are really well thought out, just something conjured on the spot, so don't take them as granted as a solution to a problem that may not even exist.

Anyway, the main reason for this post was to post the question found above.

Keli

[shimmin]

Mark:

I like the ideas, but I had an idea for a different way to do the un-penalizing factor. It ends up working roughly the same way, but one or the other may be easier to code, or more importantly easier to explain to the player in documentation.

The idea sums up as this
1. Keep the "percentage" penalizing factor. (ie, you still suffer a penalty for every 5% you advance. To make it scale right with what I'm about to suggest, though, I would make it a x1.4 penalty instead of a x2 penalty.)

2. Rather than apply all the prereqs at once, an adopting a fabulously low "base cost" to compensate, keep a moderate base cost (10, for example) and apply x2 no-prerequisite penalties _at certain levels in tech_. For example, you could advance up to 50% or whatever in engineering without engineering, but after that, the going gets tough.

Using chemistry purely as an example, a tech chart might go

Basic tech: Chemistry
10% Glassworking
20% Weights and Measures
30% Algebra (or Math 30%, if algebra doesn't make the "tech cut")
40% (some education level)
50% The Periodic Table
60% Atomic Theory
70% (Physics 70%)
80% Quantum Theory
90% Research and Development (or Commerce 85%, barring R&D's existence)
100% (Computing 105%)

This ends up working more or less the same way as your orginal proposal, except instead of compensating enormous prequisite penalties with a small base cost, it waives the penalties until a certain point of advancement. The Greeks' had an atomic theory based on Zeno's paradoxes. It didn't help them advance chemistry much; alas, for all their strength in geometry, they knew rather little algebra.

It doesn't have to be 10,20,30, either: you could well modify the above table by putting Weights and Measures, Algebra, and the education requirement ALL at the 30% level.
This doesn't "force" the player along as strict a path, but still leaves some Breakthrough techs along the way. This way, a tech wouldn't necessarily need 10 prereqs along its path; 5 or 12 would work just as well, without modifiying the basic cost (this might help for handling both "basic techs" that we were able to consistently develop throughout time (agriculture) and those the languished in infancy for millennia until the right inventions came along (computing).

[Mark_Everson]

Keli and shimmin:

What you guys are basically proposing is what I had at first. I'll try to explain the advantages of the way I'm doing it now below. The main goal for doing it the way I did is to make the tech tree more flexible than in Civ and most games in the genre. This is something LGJ is shooting for too, although I obviously can't say at this point whether he agrees with my specific implementation or not.

First Keli's questions:
This is a fairly new concept I'm springing on everyone. We'll see what people think of it. To your specific points... a) no; b) I'll discuss below, and c) that's a cute idea. But even though that is realistic, I'm not sure it is good for gameplay. It might allow stupid, unrealistic, "dodging around" strategies like you can do with the tech tree in Civ. Plus I discuss another problem with it below.

In general, the reason I decided that I wanted to stack all the prerequisites at the beginning of a basic tech is for flexibility. Let's take the case of Agriculture as a basic tech. IMO there are two very distinct strains in agriculture. There is an engineering aspect found in irrigation techniques, drainage, and plow construction. On the other side there are the Breeding aspects in getting improved grains and livestock. Either, or both of these tracks can be pursued fairly early. Obviously, if good agricultural numbers are your focus you'll need to do both. But let's look at this crude example to see why I think you can't easily restrict advance prerequisites to certain windows.

Let's suppose in the initial selection of my Civ I've picked my people to be absolutely brilliant in the engineering area, and they're pretty poor in everything else. So my agriculture can do fairly well, even if I don't meet other civs to trade techs with. I'd only be missing out on the one breeding prerequisite. If I'm isolated, it might take quite awhile before either I achieve the Breeding tech, or run into someone else that has it. But when I do, there will be a renaissance in my agricultural technology because of all the more refined plant and animal species I have to deal with. So, conceivably, I might be as high as Agriculture 30% or higher before the Breeding prerequisite comes along. However, in the normal course of things, Breeding would be a fairly early achievement, at around Agriculture 10% or so. This is why I think restricting prerequisites to a certain range will stifle the player a bit. It may be an even worse factor in some parts of the tech tree that I haven't thought of yet, things like religion...

That's my take on the issue...

Mark

[Mark_Everson]

Proposal for how to include a limited part of the technology model in demo 3.

First of all, demo 3 is very limited in scope, so there's only so much technology we could actually use anyway. What I'm proposing here it is a limited test using a few different technologies, and things you can build. Demo three will still be heavy on military stuff, because that is the first thing or working on implementing the new models. For that reason is tech proposal will also be very heavy in military-related techs and units. I have tried to include a lot of the types of interconnection's that we'll occur or in the fall model even in this simple demonstration. What I hope we can get out of it is a general feeling as to whether we are on the right track for the technology model. Does the idea of using two classes of technologies, basic techs, and "normal techs" feel right?

It's going to take some imagination on the part of the playtesters to give us a realistic appraisal. I have intentionally limited this proposal and several ways so that we can IMO do it fairly quickly, using stuff that is already in my "old" version of Clash. That means just using the old economic system even though it's not going to be around very long. I am assuming we're using the method for how the cost of basic techs is determine that I outlined recently. If everyone decide that is no good, then obviously I'll have to change the proposal . With that in mind, here it is:

Aside from the Army-specific stuff this is all from LGJ's most recent tech model (hasn't been released yet). The basic techs I propose we use for demo three are:
Agriculture (agricultural productivity)
Metallurgy
Army (a stand-in for several types of military technology)
Weapons (quality of personal weapons made in the society, I think this one specifically could be added to the tech model, and would like to hear what people think about it)

"Normal" techs used are:
Bronze Working
Iron Production
Breeding (understanding of how to influence evolution to create new plant and animal varieties that produce more or better food, etc.)
Professional Army (tech advance showing your culture knows the difference between "weekend warriors" and trained and hardened professionals)
All of these are very early civilization advances.

Now the prerequisites of each, and their weighting. Before the name is the cost, or cost basis for the tech. Basic techs and "normal techs" (can anyone think of a less lame descriptive name?) Have different scales for their costs. This is partly because the price for a basic tech is the price for a 1% improvement, whereas for a regular tech you get it all in one shot. For the demo, I propose that we start each basic tech at 5%. So, for instance, the Metallurgy prerequisite for Bronze Working is already met.

Cost Tech ___ Prerequisites (importance)
10 Agriculture: Breeding (3x); Iron Production (1.5x) [Iron is because of better tools]
20 Metallurgy: Bronze Working (2x); Iron Production (3x) [can only grow so far without these]
20 Army: Professional Army (3x); Weapons (2x)
10 Weapons: Bronze Working (2x); Iron Production (3x); Metallurgy 8% (2x)
30 Bronze Working: Metallurgy 3% (2x) [already met]
80 Iron Production: Bronze Working (2x); Metallurgy 6% (2x) [my, metallurgy and the metal-working advances are inbred aren't they? Will this be fun or exasperating?]
30 Breeding: No Prerequisites (at least in the demo version)
50 Professional Army: No Prerequisites...

Ok, but what can we Do with this stuff?
Well, for demo 3,you'll be able to build two different kinds of army units (spearmen, and legions) and four different levels of farm. Actually, your people will build the farms, and they'll automatically build the best one your technology is capable of. Legions will be significantly better than spearmen, so that once you get the chance to build legions, it will be a no-brainer that is done automatically for you. As LGJ has outlined, all this stuff you build will have its cost reduced as its associated basic tech becomes better. (I'm not sure this will work for everything we do, but we'll stick with the program for now). So, for instance, legions are (surprise!) Associated with the Army basic tech.

Each item you can build will have both take technology required to build it, and an associated basic tech that modifies its cost. In the case of the examples I've come up with, I'm embarassed to admit, the "enabling" technology and the base technology are the same. This will frequently not be true, it just so happens to be the case here. First, the military units:

Unit______Cost__Power_Enabling Tech/ Basic Tech
Spearmen__125_____4____Army 3% / Army 3%
Legion_____50_____4____Army 6% / Army 8%
(The units having the same power is caused by AI limitations I don't want to fix now. So I just made Legions a better deal at the same power, which isn't realistic At All.)

The legion illustrates the case where you can build them at a penalty before your Army No. is up to the point where they're at their nominal cost. This example is admittedly cheesy.

Now, back to the farms:

first, the old economic model is rather complicated, and I don't want to go into exact descriptions here. It will be replaced shortly anyway. So I will just given a simple sketch of what changing farm productivity through technological progress will do. It won't be correct in detail, but will give the right general idea. The old economic model will automatically upgrade farms to a new technology when it's available. A production of 1 is pretty close to subsistence level. A farm employs one "head" = 1000 people by definition. Even small increases in productivity, which you see below, can have big effects on the Civ because almost everyone is farming. If farm productivity increases by 20%, that means you could have 20% of your people building things instead of feeding themselves. It's a Big effect.

Farm Type__Cost__Produces_Enabling Tech/ Basic Tech
Farm 1_____10____1_______ Agriculture 5% / Agriculture 5% (these will all be the same)
Farm 2_____10____1.2______ Agriculture 10% (so I will only write it once)
Farm 3_____10____1.4______ Agriculture 15%
Farm 4_____10____1.7______ Agriculture 20%

Where idea points come from:

the people generate some idea points on their own. This will be fairly complicated in the real model, but handled simply here. The people in the agriculture sector will contribute some idea points to Breeding and Agriculture for free by themselves. Every time your soldiers fight, they will contribute idea points to your Army basic tech. For right now the other techs get no free idea points.

You can also buy research using money. We'll try to provide a crude interface (based on something I already have) that will give the player a variety of choices. The biggest choice will be money per turn spent across all tech. Once that a set by the player, then they'll be able to set which percent of this total pot goes into each of the eight Tech areas in the demo. If you achieve an advance, for example Breeding, it's percentage will get automatically distributed pro rated to the other remaining techs.

I, or Anton/rentgen will probably try to write a very crude tech AI to govern what the AIs do. The most likely thing we'll do is just give each AI Civ of fixed ratio at the beginning and let them play the whole game with that ratio between the 8 techs.

Mark

[This message has been edited by Mark_Everson (edited November 02, 1999).]

[Lord God Jinnai]

Originally posted by Mark_Everson on 11-01-1999 09:47 PM
cost basis for the tech. Basic techs and "normal techs" (can anyone think of a less lame descriptive name?)

Tech Advancements?

Where idea points come from:

You can also buy research using money.

Careful on this one...it should be used with a dimishing returns graph and to the point where any additional money spent is useless since some techs can't be done overnight no matter how much $ u throw at it.

I, or Anton/rentgen will probably try to write a very crude tech AI to govern what the AIs do. The most likely thing we'll do is just give each AI Civ of fixed ratio at the beginning and let them play the whole game with that ratio between the 8 techs.

I'd still like to try the later versions once i learn java more.

Mark

[This message has been edited by Mark_Everson (edited November 02, 1999).]

Anyway now to my stuff.

I am trying to make technology r&d flexible and do not like the tree method which to me seems to promote to much inflexiblity... although it might seem good for the demo, i think the final product would be much better without it. This is because there may be more than 1 way of approaching something as well as the many 'soft rq' and the basic technology 'soft rq' policy...also Mark, and anyone else, i haven't heard anything about what should be good percentage decreases/increases per % point since i favor a stricter one since many techs will require low # on basic techs and thus u could know nothing about it and still research it.

The one way around that i see it a min of 1% add-on line somewhere in the coding, however i still think it shouldn't be too gradual.

[Mark_Everson]

Hi Lord God Jinnai:

"Tech Advancements"... better than what I came up with

>>You can also buy research using money.

>Careful on this one...it should be used with a dimishing returns graph and to
>thepoint where any additional money spent is useless since some techs can't
>be done overnight no matter how much $ u throw at it.

Agree completely; this is just for the demo. There should be intertia so that slow but steady funding beats thrashing back and forth every time. There are some suggestions on how to do this in the Civ3 List Tech Thread, but I think its pretty straightforward.

>I'd still like to try the later versions once i learn java more.

Don't think that's a problem. What parts would you like to do? I'm sure we can work something out.

Rentgen: Is there a stopping point on tech you'd be comfortable with?

>inflexibility of demo tech tree.
I didn't put any either/or stuff in... I know we plan to have that in the real model. Is that what you mean?

Cost advantage/hit per % diff in a Basic Tech...
I think we can try your earlier suggestion of 10% advantage per point in basic tech on the plus side (up to 50%); and double that on the downside. IMO its easily tuneable or made nonlinear, so lets just wait and see how your first suggestion works.

Mark

[rentgen]

Originally posted by Mark_Everson on 11-02-1999 08:45 PM
Rentgen: Is there a stopping point on tech you'd be comfortable with?

--- not sure i understand your question correctly... Do you mean am i willing not to code the tech part ?
If LGJ has his own model and actually wants to implement it, why not? I can just quickly do what was proposed by you earlier for the upcoming demo...

-- Anton

[Mark_Everson]

Rentgen:

Yes, As in set up a good class structure as we are discussing and implement what needs to be done for demo 3 and perhaps somewhat beyond, and then switch to something else.

However, Only if its OK with you... We're doing everything by mutual consent.

LGJ:

There's still a Huge amount of work involved in defining the model. When do you think you'd be up to speed enough in java to take it over? Alternatively, would you like to work on the the GUI for Tech and leave the 'guts' to Rentgen? Because once the model is speced out and the main classes are done, there's just not that much else to do other than AI and GUI.

Mark

[rentgen]

Originally posted by Mark_Everson on 11-03-1999 06:27 AM
Rentgen:

Yes, As in set up a good class structure as we are discussing and implement what needs to be done for demo 3 and perhaps somewhat beyond, and then switch to something else.

I wrote such a post and due to a silly feature of the cgi program that handles this it got lost!!
Anyways, i was saying that i absolutely don't mind. And then i started a long philosophycal talk about design issues i have in mind. Here it is in a shorter form.

Fisrt of all, this is kind of raw material.. and i didn't plan to post it for a while but issues started to rise. This is a short description of what i'd like to do with a tech package now and it's possible application in other parts.
The idea is simple: use interfaces. I'll try to define interfaces for the tech package that are to be used outside of it by the rest of the code. Then i'll provide a simple implementation of these intefaces for the demo based on what's been discussed. The advantage of this approach is huge: when you, or LGJ or me or anyone else wants comes up with a different model he just needs to implement the interfaces the way he likes -- and voila! No changes in the rest of the code are needed at all. Not to mention ease of support and future development.
That's why i was telling Mark it would be good to have a thouroghly thought out design model and only then proceed with coding. However, he convinced me this was impossible -- due to the impatient nature of human beings
If this works (and i really hope so) i hope we can redesign in the same fashion the rest of the existing code, which will make it more flexible and as such easier to maintain, add features etc.
Just some quick ideas...

-- Anton.

[Mark_Everson]

Rentgen:

My sympathies, I Hate it when my hard-written posts get eaten. I have taken to writing long ones in a word processor, and then pasting them to protect myself.

Thanks for not minding us playing coder volleyball with you

I think your idea of using the tech model as a demo for your interface proposal is good. I suspect you're right that its the best way to go. If you re-write your interface suggestions for the project, probably you should put it in the coding issues thread.

[Lord God Jinnai]

It seems like it might be better to wait for the GUI since no one seems to want to go to the library and pick up my book (I can't legally drive do to my eyesight). It might also be a good idea if i help in editing instead of coming up with the original coding.