Technology Tree

[Anon Zytose]

Due to recent discussion in the Cottages thread, I think it's about time this got its own thread.

I personally see the technology tree as the backbone and foundation of pretty much any of these turn-based empire-building games. The extent of what's possible grows from it. And if there's anything I can get myself to focus on in this mod, it's this.

One step to take is to determine every application and tech-derived bonus to exist in the game, and get an idea of an order of when they may tend to show up.

Another step is to figure out what sort of technology would go into each of the facilities and projects and unit components and acts of terraforming, then place same-era applications involving the same kinds of technology into small groups (mostly pairs). In many cases we'll need to define and understand each and every tech advance to go into the game.

Then, of course, there's the effort of organizing all the tech advances so they have reasonable prerequisites, tend to show up at reasonable times, and so that the average number of advances available at any given time is generally some comfortable value like 6.

This post is probably still under construction.

 

[Anon Zytose]

Anyway, one major component to bring to attention (that I missed last time) is making sure there's enough stuff to go into each advance. So that on average two advances of the same cost will be equally worthwhile. This (plus how early and late techs no longer cost the same) will provide a sure step-up from when I'd rarely discover Photon/Wave Mechanics before Eudaimonia and Secrets of Alpha Centauri. ...For that record, it might also help if most tech advances provide either a civilian application of something to help to keep mind worms at bay. We can have some civilian applications have multiple tech prerequisites each if necessary (and makes enough sense). And have the pure military techs provide benefits that actually require that tech. As opposed to going for Missile Launchers and forgoing Superconductor and Nonlinear Mathematics until one discovers Cyberethics (or maybe Advanced Subatomic theory).

Yeah, I know I'm pointing at SMAC/X and not Planetfall. But I have yet to check where all the applications in Planetfall go. It'd probably be easier if I had the second BTS disk with me, but I guess I'll just make do with what I have for now.

Anyway, there are still definitely plenty of things with Planetfall I'd like fixed as well: Squeeze it back to where the number of advances at any one time is mostly six (until near the end, at least). Add more tech advances (and their applications) in total. Do something about poorly defined advances. Plus still keep new ideas rolling in, like greenhouses. ^^ And yes, I'm seeing what I can do about all that right now. I currently have a file with some 82 entries and some possible applications in each. However there might be some holes in terms of secret projects, and figuring out what weapons to place in the game and where is in many cases more difficult than I expected. But I'll try and figure stuff out.

And input is welcome.

 

[The Beard]

Is there a way for us to implement an option for Blind Research? I used to not like it so much, but now that I have SMACX, I do like the contrast between the humans' inability to know exactly what they are researching, and the Progenitors knowledge of what exactly needs to be researched. I can't imagine the Blind Research thing would be too difficult to program, but we'd have to make sure that the computer is smart enough to make "good choices" for the techs we research. Or if they are just random/dumb, then all of the factions have to have the same dumb AI "choosing" what they are researching. And then Espionage should only reveal the concentration of what another faction is researching, Only espionage against Progenitors would yield what specific techs were being researched (also, it should be much harder to spy on cultures alien from you own, if only for the simple fact that no one that you have looks like anyone they have. It would have to be done entirely through bribery and computer hacking, for the most part.

 

[Anon Zytose]

I personally have no idea how to model Blind Research. I'm also not familiar enough with it to not be sure if someone who knows all the priority values could cheat the system. For both of these, you may need to ask someone else. ^^;

 

[Maniac]

Good thread idea. There's a 200+ post thread on the tech tree in the private forum, but most of it contains outdated discussions, so I didn't see the point in making it public.


Anyway, one thing many people have said is that there are too many tech choices available at any particular point in the game. This could be easily solved by changing many of the OR-prerequisites in the tech tree into AND-prereqs.

However personally I'm not convinced it's a bad thing you have so many tech choices. So I'd like to ask the question: WHY is it good to only have six choices on average? WHY is it bad to have more?


As far as I can see, the advantage of the current tech tree with lots of OR-prereqs is that it allows a wider variety of possible strategies.

The drawback is that it is confusing for new players of Planetfall. They can't see the forest through the many trees. Currently you guys are all Planetfall newbies, so it's normal you feel overwhelmed. Can't be sure of that yet of course, but my expectation is that as time progresses and you become more familiar with Planetfall, you'll better be able to decide at the blink of an eye what are the important tech choices available to you for the strategy you're following. At least I can.

So I think the basic question here IMO is, should the game be designed with new players in mind, or should it be designed to offer a wide variety of possible strategies, so that the game is still fun to play for veterans in say two years?


Btw, have you guys checked out TechTreeManiac.xls which comes included in the Planetfall download? Perhaps the colour coding I use there might shed some more light. I've attached the latest version here.

 

[The Beard]

Well, part of the problem of the current tech tree is that it is hilariously easy to take advantage of. While I was indeed totally lost on the first round of stuff, the second time I tried playing, I was unstoppable. The fact remains that it is possible to research Superconductors before High Energy Lasers, which gets you a strength 5 weapon before a Strength 2 weapon. There are so many other techs that the AI doesn't really know what to do, and just goes through the techs "in order." So I researched Superconductors as one of my early techs, and destroyed the Hive and the Morganites in about 10 turns, and then went searching about Planet to find everyone else. Even though the timespan was quite long, no one put up much of a fight, as they did all of the combat techs "in order."

I think, at the very least, that weapon techs should be for the most part strictly linear. There is some wiggle room with some of the later weapons (like who cares if you have strength 14 or 13 rovers?), but the ability to get such techs so early is ridiculous--they only take about twice the time to research as the other techs, and are clearly worth the wait.

A narrower choice in your tech tree prevents people from taking advantage of the tree, and also makes it easier for the AI to compete.

 

[Anon Zytose]

Thanks. ^^

I haven't checked, but if we turned ALL of the Or-prerequisites into And-prerequisites, we'd probably have a tech tree with an average number of choices most of us are used to. (Probably around 5.)

If this much variety is to be available, all those choices will need to be worthwhile. And I do mean ALL of them. If we find that nearly all the time (like over 99%) that it's more worthwhile to research A before B than B before A, then either the advances need to be rebalanced. If we can't do that, B may as well require A. (It kinda makes me think of how much I seek out mind worms due to their strengths as early game units against pretty much anything.)

Civilization IV remains fun for me after two years (part of why I wish I didn't forget my BTS disk 2 at home while reinstalling everything), and the average number of tech choices (at least for Vanilla and Warlords) is this many:

Spoiler :

Or do I suck at the game too much (I have trouble with Noble) to count as veteran? o.O

Somehow I'm reminded of one time some years ago I thought of an idea of giving Alpha Centauri some options on how broad or narrow the tech tree could be. This was soon followed by the idea of giving tech advances random prerequisites (like Probability Mechanics might require Pre-Sentient Algorithms or it might just require Cyberethics). It didn't go very far, of course, but I do wonder now what would happen if someone modded Planetfall to have two or three separate tech trees, one of which would be chosen at the start of the game.

 

[Anon Zytose]

Well, part of the problem of the current tech tree is that it is hilariously easy to take advantage of. While I was indeed totally lost on the first round of stuff, the second time I tried playing, I was unstoppable. The fact remains that it is possible to research Superconductors before High Energy Lasers, which gets you a strength 5 weapon before a Strength 2 weapon. There are so many other techs that the AI doesn't really know what to do, and just goes through the techs "in order." So I researched Superconductors as one of my early techs, and destroyed the Hive and the Morganites in about 10 turns, and then went searching about Planet to find everyone else. Even though the timespan was quite long, no one put up much of a fight, as they did all of the combat techs "in order."

I think, at the very least, that weapon techs should be for the most part strictly linear. There is some wiggle room with some of the later weapons (like who cares if you have strength 14 or 13 rovers?), but the ability to get such techs so early is ridiculous--they only take about twice the time to research as the other techs, and are clearly worth the wait.

A narrower choice in your tech tree prevents people from taking advantage of the tree, and also makes it easier for the AI to compete.

Makes me think of how all the techs in Vanilla SMAC that provide armor all lines up.

I was mainly thinking there'd be some seven generations of weapons (and armor). Each generation would have all of them overall even but with various bonuses against certain kinds of chasses, armor, or each other. The first would probably just be conventional rifles, no armor, and hatchling psi (usable by Planet only). While the last generation would be filled with singularity lasers, stasis generators, warp bubbles, string disruptors, graviton guns, tachyonic missiles, demon boil psi (both as weaponry and shielding), and a few other things.

Will we have to wait on wider tech trees until we get Digital Consciousness for real, or would it then be narrow enough so that unmodified humans could keep up?

 

[Maniac]

Well, part of the problem of the current tech tree is that it is hilariously easy to take advantage of. ... The fact remains that it is possible to research Superconductors before High Energy Lasers, which gets you a strength 5 weapon before a Strength 2 weapon.

Extract of PlF v4 changelog:

26. Gatling Laser now strength 4.

Anyway, because the tech tree and content is being juggled around all the time, I haven't spent any time yet on trying to improve the AI. So yeah, the AI just picks random techs to research at the moment, and they're kind of a pushover at the moment. It's definitely to improve AI behaviour though, so the situation you describe wouldn't happen anymore.

I think, at the very least, that weapon techs should be for the most part strictly linear. There is some wiggle room with some of the later weapons (like who cares if you have strength 14 or 13 rovers?), but the ability to get such techs so early is ridiculous--they only take about twice the time to research as the other techs, and are clearly worth the wait.

Personally I'm more hoping military benefits could be spread all over the tech tree. Meaning while you spent time researching Gatling Lasers, someone else used their time getting eg artillery, needlejets or mind worms, and could hold you off.

If this much variety is to be available, all those choices will need to be worthwhile. And I do mean ALL of them.

Yeah, that's my hope/ambition. I'd like to give it a try reaching that ideal before 'giving up' and putting lots of AND-prereqs in the tech tree.

Civilization IV remains fun for me after two years

Personally I find the unmodded Civ4 game boring because there is so little variety in research paths. So I definitely need more variety than the average player to find a strategy game fun.

 

[The Beard]

That's good about the change of Gatling Lasers.

What I meant about having the weapon techs be linear, is just so that we avoid having gatling vs. guns, Tachyon Bolts vs. Gatling and whatnot. I mean, it is good that once you research a new weapon that you get to kick everyone's ass, but I would rather have there be a chance that the AI will actually be able to compete with me, rather than me remaining undisputed warlord of Planet for hundreds of years. So yes, what you said about the chassis and whatnot is still valid--it is just the weapons I want locked into their own "eras." It is a bit embarrassing when I have a 95.34 percent chance of completely obliterating units defending a city, even when it has its defenses intact.

Another thing--I realize that the mechanics of CIV IV make the armor/weapon distinction of SMAC a bit blurry, but is there a possibility of having the "armors" be automatic bonuses for defending? I guess this would make these units take the place of "archers." like for example, Synthmetal Armor would be +1 strength when defending, Probability Sheath would be +5, Stasis Generator would be +11, etc. This way there could be a lot more "universal benefits" for techs. If it isn't possible to give discrete values as bonuses, I suppose that the armors could be employed as +percentage points, but that could be a bit unbalancing towards the end, as I doubt that a stasis generator would help a singularity laser hovertank very much, but even a +5% or +10% could be unbalancing (or at least not really make sense, as synthmetal would certainly be ripped to shreds by any strength 20+ weapon anyway.

 

[Anon Zytose]

One thing I've been thinking about strength vs. att/def. If army A attacks army B, army A's weapons have to get past army B's shielding. If army B's shields are really strong, it'll slow down army A. However, at the same time army B needs for its weapons to get past the shields of army A, or else army A will eventually wipe out army B anyway.

I figured that the strength of a unit is essentially W*D, where W is the effectiveness of the weapon and D is the effectiveness of the armor or shield or whatever. For example, if one unit with W and D of 12 each fights a unit with W and D of 1 each, the first unit kills at 12 times normal rate while the other unit kills at 1/12 normal rate. The first unit is fighting 144 times as well as the second unit.

Now in Civ4, not counting scouts or animals, the strongest unit has 20 times the strength of the weakest combat unit (if 40 is the highest strength value). We may want it to be where the last-era weapons and armor have roughly 20^.5=4.47 times the strength multiplier as the conventional rifles and lack of armor. Like maybe ranging from 4 to 18.

 

[The Beard]

Although, realistically, I think with a name like "Singularity Laser," that there should be very few defenses against it. Synthmetal or no, it should cut through it like butter. The only things with much of a defense would probably be neutronium, stasis generators and the like. Part of the fun of getting into the end game of Alpha Centauri is that all that is left are your most dangerous opponents. Once you get the Singularity laser, you can safely declare war on capitulated states and on weaker allies in favor of having a final battle between you and a similarly endowed faction. Whipping out behemoths and sky fortresses in the endgame is glorious.

So I'm in favor of a bit wider gap than 4 to 18.

 

[Anon Zytose]

Well, with a range of 4-18, no armor would be at 4, synthmetal would be at around 5 (not 6), neutronium would be at around 13, and the stasis generator would have the top at 18. Similarly, those conventional rifles would be at 4 while a level-1 weapon would be at 5 and the singularity laser would get to 18. So a scout patrol would be at 16, a synthmetal sentinel would be at 20, a synthmetal unit with the first laser would be at 25, and someone with a stasis generator and a singularity laser would be at 324.

My main thought of combat is that regardless of who strikes first, both sides are trying to slice through each other. While the singularity stasis troops pretty much would slice through synthmetal garrisons like butter (much like a tank vs. an axeman), they might not find it quite so easy if their weak-armored opponents also have singularity lasers and are slicing back at a decent rate. The stasis/singularity unit would still win, but it'd be 324 vs. 90. Kinda like Modern Armor vs. a Grenadier, or Cavalry vs. a Chariot.

Of course, I have thought of how in SMAC/X, the relative strengths of the most powerful non-psi units would tower at 48 or 96-120 times that of a fission-powered scout patrol, depending on who's attacking. Thus providing a geometric mean of around 72. Assuming the no armor and conventional rifle components remain at 4 each, this would raise the multipliers of the last-era weapons and armor to 34 each. I guess the next question is if the relative strengths of the units have any business rising ln(72.25)/ln(20)=1.43 times as quickly as the ones in regular Civ4.

Maybe we should place weapons and armor into its own thread?

 

[The Beard]

Although, I do think that the multiplication of Weapon and Defense levels is a bit drastic. In my mind, having defense is something more like a percentage multiplier, or an additive operation. That way we can do multiplication by percentage points with the various reactors, and not end up with anything like 1080 vs 20. Because I think that if things get too big, then the end-game units will waste any previous levels, even with base defenses taken into account.

 

[Anon Zytose]

I don't see how multiplying armor and defense into strength would be drastic, as long as the final products are reasonable levels of power in relation to each other. Like Singularity/Stasis going double against Shard/Neutronium going double against Chaos/Photon going double against Gatling/PlasmaSteel going double against Scout Patrols. I guess I can see that the products might look a lot more daunting if ranging from 16 to 324 (as opposed to 2 to 40.5). If necessary (and if we can use non-integer strengths), we can drop the ranges to, say, 1.4-6.3, for strength ranges of 1.96 to 39.69.

I had thought of making defense as a percentage multiplier. (A universal strength multiplier at that.) However, toward the end of the game, the percentage increase would have to rise to something like +347%. If I'm right about this, it would greatly weaken the relative effect of other bonuses like Comm Jammers vs. fast units or city defense for fortified infantry or whatever.

If we ever do have anything like 1080 vs. 20, it's because some of the strongest units in the game were pitted against some of the weakest combat units of the game, AND we ended up deciding the strongest weapons and defenses should be at least some 7.35 times as effective as those of the scout patrols. If the range s 4-18 and we get 324 vs. 16, it'd be much like a warrior fighting against a modern armor unit in normal Civ4. Similar odds toward who would win, and very low odds toward the battle happening at all in a particular game.

I personally think one reason units strengths in Civ4 multiply by 20 over the course of the game (instead of 72 or whatever) is because they want the units of one era to have some chance at defending against the units of the next era, at least if they build enough of them and mostly hold their guard in their cities.

 

[The Beard]

Suppose we have a situation where the attacking party has a powerful weapon and good armor, and the defending party only has a weapon of strength one, but a defense of many multiples of the strength of the attacking party. Is it possible for us to have something which simply renders the attack unsuccessful? (the attacking party simply retreats)

Because that was one thing that never made sense in SMAC--when I attacked with a gravship with a stasis field and singularity laser, it shouldn't be possible for me to be killed by someone with a strength 1 gun and stasis field armor.

I'm not sure how we should go about this--if we only have the armors go for the defending party, then any attack in Planetfall is likely to end in failure. Unless we somehow make it prohibitively expensive to have a high armor value and a high weapon value--but I'm not sure that is possible with our current mechanics. And if we have the defense be an added value or a multiplier, then armor is simply acting like a new weapon, and there isn't any point in putting it in the game at all.

Perhaps if we could have a defense strength from the armor which only is used when defending, and an attacking strength which is only from the weapon when attacking? In effect this would be simulating SMAC's system within CIV IV's "simplified" system.

Otherwise we could simply state that weapons have become so powerful that mobility is much more valuable than armor, (such as how muskets made suits of armor obsolete), and that there is almost no point in wearing it if your enemy has a tachyon gun.

Also, this should probably be put in its own thread I suppose. It is probably an issue which needs a good bit of discussion.