SMAX - The Will to Power - mod

[bvanevery]

Such SPs give just pure economical advantage.

Some give logistical or military advantage, when building or conquering a base.  For instance if I have the Command Nexus, and I build a sea base, I don't have to worry about building a Command Center there, before rushing a defensive unit with some training.  A Perimeter Defense upon conquest is definitely useful.  Naval Yard and Aerospace Complex give defense bonuses, so having them everywhere automatically is useful.  I don't have to spend extra turns rushing a defensive building.

Ok who's gonna hack up the Punishment Sphere Every Base SP?   

[Hagen0]

I think you are bit defensive about this, Tim. Of course, there will always be tech progression that makes little sense or is hand-wavy at best. The original tech tree is, however, much more reasonable than the current one of the mod. The majority of the prerequisites is plausible in vanilla.

Would you be OK with a counterexample? I don't want to presume. It's your mod. What I noticed when thinking about it was that interleaving the early combat techs with builder techs has its downsides. The consequence is that you can can pursue builder techs as usual and get up-to-date military capability also because it is required.

One could require Conquer techs have their own path like bvanevery has it. Or you could give Conquer factions access to earlier stronger weapons in their tech path and add some weaker and later options for builder. For instance level 3 Superconductor for 4 power weapons in the red path and level 4/5 Nonlinear Maths in the builder path for 3 power weapons.

[Nexii]

I considered making the armor techs and defensive unit ability techs DISCOVER. CONQUER would get the weapons techs, chassis aside from Foil, and the offensive abilities. EXPLORE and BUILD have a lot more units/facilities/etc.

I suppose the next step is to look at all the facilities, SPs, SEs, and unit options and categorize them similarly.

[lolada]

This is nice talk in theory, but in practice we want AI to be good and for that AI needs access to good facilities and weapons. If its so directed that one AI goes for explore and almost never gets any weapons.. it could be piss poor AI that dies.
Also these conquer aggressive AIs are already quite terrible at infrastructure building.. or teching economy..

So the tree should be somehow setup to help AI get good performance.. player can manage.

[Alpha Centauri Bear]

I considered making the armor techs and defensive unit ability techs DISCOVER. CONQUER would get the weapons techs, chassis aside from Foil, and the offensive abilities. EXPLORE and BUILD have a lot more units/facilities/etc.

Armor is both for conquest and protection from it. So is weapon. The defender should use a lot of attacks, bombardments, espionage, etc. to have effective active defense. And, therefore, attacker should armor their units, etc. Maybe the ratio could be slightly different. Like assailant could weapon/armor 50/50. Whereas, defender may emphasize on armored defenders in base for 60-80%. However, they need both weapon and armor. It is kinda difficult to conquer with contemporary weapon and [poopy] armor.

Again, don't be bought by the color/designation of the tech. Each has 4 values for each priority. The 4333 one could be conquer but it is practically everything. It is more explore than 0001.

[Alpha Centauri Bear]

I think you are bit defensive about this, Tim. Of course, there will always be tech progression that makes little sense or is hand-wavy at best. The original tech tree is, however, much more reasonable than the current one of the mod. The majority of the prerequisites is plausible in vanilla.

Not at all. I don't think in term of paths. Practical gaming shows there is no isolated development. Everybody trades about half of techs. Plus high randomness with blind research. So one would research about 30-40% of what they like and the rest of it will be traded or randomly picked by blind research functionality.
The research priority let you research certain techs somewhat earlier but they do not let you run away researching only certain color. Probably only 1-2 levels ahead.

I would yet to see a tree designed with paths in mind. It would be interesting to study it.

My tree also links weapons and armor but I don't consider them a "path". It is just a prerequisites to make sure one don't get military items in order. Your path theory would predict pure conquest oriented factions run away in weapon and armor tech far far ahead of others then. That doesn't happen in practice. They are ahead by 1-2 levels comparing to other non-conquest faction with equal number of discoveries. And this gap stays the same throughout the game.

[Alpha Centauri Bear]

This is nice talk in theory, but in practice we want AI to be good and for that AI needs access to good facilities and weapons. If its so directed that one AI goes for explore and almost never gets any weapons.. it could be piss poor AI that dies.
Also these conquer aggressive AIs are already quite terrible at infrastructure building.. or teching economy..

So the tree should be somehow setup to help AI get good performance.. player can manage.

Exactemundo, my friend!
Everybody needs everything or almost everything. Some faction may skip on some features here and there but majority of it are equally useful for all. People talk about research paths as if they want to give one player all weapons and another all facilities. The difference in factions is their behavior not the tool-set. You need pretty much everything for assault as for protection. There are just few things needed for specific purpose. Like PD is for protection only. Everything else is needed for everybody else: formers, weapons, armor, almost all abilities, 95% of facilities, etc.

I would encourage everybody to not look at tech single designation (color) but on its weights (4 numbers). The higher numbers in certain area will make this tech a desirable pick for some factions with corresponding priority. I have already adjusted them to aid AI who stick with their priorities for the whole game. Human can maneuver.

So just because human changes their priority often there is no point in having set paths in tree spanning from beginning to the end. Human will just jump between them as needed. The biggest problem of balancing the tree is to help rigid AI with research. The difficult part is not to give AI what they want according to their research priorities but to give them what they do not want! 🤣

Indeed I've observed this in almost every game. AI it fine pursuing their path. They never can run too far ahead in it but they constantly have some small advantage edge. At the same time they can seriously fall behind in essential stuff outside of their focus. And here is where we need to help them. Make some essential features accessible to everyone.

[Alpha Centauri Bear]

I have computer model to see how different research priorities affect the development path and to make sure they are not too much skewed. Feel free to propose any improvement in this regard.

[Hagen0]

Interleaving the builder and military techs is fine with me. I was just asking about the priorities.

Designing a tech tree is really hard with your requirement that each tech is prerequisite for one tech in the next tier and one in the tier after that. Each tier of techs has a one-to-one relation to 4 other tiers. That is amazingly restrictive. Perhaps you could relax the requirement for each tech in tier n to have a prerequisite in tier (n-1) and another in one of tier (n-1), (n-2), (n-3) with n-2 strongly prefferred?

[bvanevery]

Or you could give Conquer factions access to earlier stronger weapons in their tech path and add some weaker and later options for builder.

It's not a bad idea.  What I found outrageous in the original game, was that the Gaians were getting one of the most powerful early weapons from E6 Synthetic Fossil Fuels.  Like are you kidding me, Gaians??  FFS why?

I did not object nearly as much to the Morganites gaining strength 4 armor.  I think sprinkling a modest defensive improvement is not nearly as much of a sin as sprinkling a powerful weapon.

The University getting a Chaos gun kinda bugged me, but at least it's in character for mad scientists to get powerful weapons.

[Nexii]

I considered making the 2nd prereq always n-2. I'd have to map it out and see what implications that has, if it's too restrictive. It would mean the T2 techs only have one prereq which is probably ok.

The idea for discover getting armor was more that I don't want 50% of the tree to be red techs. And yea it wouldn't mean they aren't a priority for conquer. They could be like 3,4,3,3 or the like. It's not too much of a stretch, you aren't going to conquer the world with high armor alone. Whereas with weapons and speed you could.

So I started from the other direction as well. For everything a tech can make available, categorizing as conquer, discover, build, or explore. Not surprisingly conquer had the most at 57, discover 46, build 50, and explore 51. This with the view that terraforming is build, defense is discover, anything growth/drone control is explore. This isn't so far off that a tech tree couldn't be a quarter of each. Just means putting a few more benefits in each conquer tech.

Roughly 204 tech grantable abilities and 86 techs means on average a tech will give the ability to do 2 new things. With some granting 3. This I found a bit interesting since it felt like many conquer techs in the stock tree only allowed one thing like a weapon upgrade.

And yes Synthetic Fossil Fuels is definitely a red tech

[bvanevery]

This is nice talk in theory, but in practice we want AI to be good and for that AI needs access to good facilities and weapons.

You are risking genericizing all factions to playing exactly the same way.  This is already a problem in straight Thinker Mod.  The AI abuses boreholes and condensers, forcing a human player to fight fire with fire to keep up.  You might not complain about the fiction of most factions, but for Gaians it's completely tone deaf.  They're supposed to be able to compete by having a viable alternative asymmetric tech tree.  This takes careful design.  In the original game, giving the Gaians an easy strength 6 Missile Launcher was a design cop-out.  Whatever they had imagined they were going to do with AI mindworm play, it clearly didn't work in practice.  So they activated an escape hatch where the Gaians get a more powerful weapon than they deserve.

My own tech tree isn't just for continuity of narrative and lore.  It has been subjected to lots of playtesting over 2 years.  Things are as they are because the power of factions and AI performance in stock is mostly working. 

Playtesting is in fact why development has taken 2 years and not all gotten done sooner.  There has to be a certain amount of time lag of playing games to see what happens, getting the occasional player feedback, and making design adjustments accordingly.  So many of my mod releases, I thought I was done and had covered everything!  Somehow, I've kept finding dusty details to squeeze yet more performance out of the game.  I keep thinking I'm really a lot closer to done now.  Yet, 1.43 is a more substantial revision than usual.  Well if I've calculated right, at least I'll have finally put "the satellite problem" to bed.

The biggest AI performance problem I've noticed lately is that indeed, for 2 factions a pure Conquer focus was not working.  In 1.43 I'm making a big change of diversifying faction research thus:

Explore - Cult
Discover - University
Build - Morganites
Conquer - Spartans

Explore, Discover - Gaians
Explore, Build - Pirates
Explore, Conquer - Caretakers
Discover, Build - Angels
Discover, Conquer - Usurpers
Build, Conquer - Drones

Explore, Discover, Build - Peacekeepers
Explore, Discover, Conquer - Hive
Explore, Build, Conquer - Believers
Discover, Build, Conquer - Cyborgs

This does risk genericizing AI play.  However it's a more sophisticated take than having all of them implement exactly the same strategy.  When facing random opponents, there is now likely to be a coverage of different strategies.  One immediate observable effect is there's more Discover research in general, which heats up the rate of tech research.  I don't know yet whether I'll have to rebalance anything in the Discover tree to compensate.  It could take several releases to reach a verdict or maturity on that.

The only reason these strategies have any diversity in my mod, is because I've been working towards more and more strongly separating Explore Discover Build Conquer in my mod.  If they actually did very similar things, then diversity wouldn't matter nearly so much.  How much diversity to implement, is a matter of design.  I have a strict pure Conquer policy on weapons and armor, and I will never change that, having pursued that kind of design to its logical conclusion.  I do concede now, that it's not the only valid weapons and armor design.  The original game was goofy though and did many things wrong with weapons and armors, such as giving away powerful stuff inappropriately.

[Nexii]

Yea it's pick your poison. Thinker with ICS/Adv terraforming. Which really I'm not a big fan of that style either. Or base AI that honestly plays quite bad with any sort of Conquer focus. You have to give it 2 other priorities with Conquer to make it build any infrastructure, and even then :/. I found that once I fixed the AI from tanking energy, builder factions win almost all of the games. Conquer often get a continent or so but they aren't co-ordinated enough to break a fortress most of the time. Sometimes with lots of drop troops they can. Single continent maps I give conquer factions much better odds, as long as they aren't huge maps.

[bvanevery]

It is kinda difficult to conquer with contemporary weapon and [poopy] armor.

For the human player it's trivial.  Build rails.  The AI doesn't know how to do rails based conquest, so it's at a disadvantage.  Rails is a tried and true method as far back as Civ 1.

Some humans do the exploit of building a Truck unit, to drop off unarmored units that are going to fire.  I'm not into that kind of abuse.  Again, the AI doesn't know how to do it.  I did try implementing a Truck predefined unit, but the AI had no clue.

I have been known to use armored Drop Transports once I have orbital insertion capability.  Technically they could be used much earlier in the game as a kind of paratrooper, just at a much more limited range.  Range could in principle be extended with Carriers, but I almost never do Carrier based combat.  Yet another area the AI doesn't understand, too exploity for me.

[bvanevery]

Practical gaming shows there is no isolated development. Everybody trades about half of techs.

An unfinished SP can be used to block the AI from trading a tech with anyone.  My tree is designed around this.  SPs block the most important techs.  All the players can trade everything just fine on Tiers 1 and 2, so everyone gets their basics and Politics and Economics choices in a pretty reasonable period of time, even if separated by great distances.  But Tier 3, the tech trading door slams shut.  You simply aren't going to get "stuff" anymore.  You may say "half" of techs get traded, but it matters very much which half.  Techs are not all created equal, and SPs are the means to block them from spreading too fast.

I would yet to see a tree designed with paths in mind. It would be interesting to study it.

This is of course my mod, since forever, since the very beginning.  In 2 years you really haven't checked it out in any great detail?  The differentiation between paths has only gotten stronger in the past year.

If you mean you're looking for a mod that separates tech tree portions even more strongly than I did, well I don't think that's real world.  There's only so many techs.