Ideas for developing three of Thinker's new features

[Michael2589]

Three ideas for some of the new features

For new_base_names
Current behaviour:
Vanilla:
If it's the faction's first base, X. If it's a land base, Y1.
If it's a sea base, choose a name from the faction's #WATERBASES list of sea base names. If all of those have been used, choose a name from the faction's #BASES list of base names. If all of those have also been used, use the vanilla base name generator to create a name.
Thinker 2.8:
If it's the faction's first base, X. If it's a land base, Y2.
If it's a sea base, choose a name from the faction's #WATERBASES list of sea base names. If all of those have been used, use Thinker's base name generator to create a name.

How I'd like it to work:
If it's the faction's first base, X. If it's a land base, Y2.
If it's a sea base, choose a name from the faction's #WATERBASES list of sea base names. If all of those have been used, choose a name from the faction's #BASES list of base names. If all of those have also been used, use Thinker's base name generator to create a name.
Comment: This is both consistent with the original game's behaviour and means that you will get further into the game with flavorful base names.

Game setting: rare_supply_pods
You could mix in some monoliths among the nutrient/mineral/energy bonus resources. Monoliths are an interesting terrain feature, quite nice at the start of the game, and definitely not being transported from Earth on the Unity. I also think you could mix in some fungal towers, but that's just me. Remember, the vanilla supply pods aren't always good.

Game setting: revised_tech_cost=1

1. "Tech costs scale relative to the level of current research item, instead of the old version where costs increased according to the amount of techs a faction had."

I agree that this is the better choice when playing with directed research, but when you are playing with blind research the number of techs the faction has works better than the level of the tech being researched. (If I am playing with blind research, and I change my research priorities, then maybe the number of years expected to research the tech changes dramatically, which, besides just being weird, means that I can make a pretty good guess of what the tech is.)

Note that Progenitors always get directed research even when the game rules are set to blind research.

2. from details.md: "* Count every other faction with commlink to the current faction who has the tech, while infiltrated factions count twice. Discount 5% for every point in this value. Maximum allowed discount from this step is 30%. Only infiltration gained using probe teams counts for the purposes of extra discount, the Empath Guild or Planetary Governor status is ignored in this step."

Alternate choice: Factions that are either infiltrated or are in a Pact with the current faction count twice.

Comment: Historically, there are a lot of examples of allied countries sharing their technology.

[induktio]

Okay, let's see. :)

I'm not sure if I follow you with the base naming description. Thinker's new_base_names is supposed to use the same vanilla naming scheme until it runs out of names in the faction definition, and then generate new unique (generic) names for new bases. So not sure if there's a difference in the beginning.

About supply pods, it's certainly possible to add more monoliths on the maps. This can be done with the new world builder code. In fact, because the supply pod frequency is reduced, fewer monoliths are spawned on the map, so this could be adjusted closer to the original frequency. Maybe change like one fifth of the resource bonus tiles to monoliths.

Also it's true revised_tech_cost is more aimed for directed research games. I'm not really fan of the blind research mechanic as it randomizes gameplay way too much and also most multiplayer games leave it disabled as far as I know. Also I have a bias against the Progenitor factions because they don't fit very well thematically with the original game lore. So I'm not really trying to balance them as first priority. But yeah, it's true revised_tech_cost attempts to model some cross border "technology transfer" with those discounts. It always happens with some degree between nations, whether allied or not. Have to think about adjusting that bonus.

Also as advance notice, I'm planning to deprecate the old terranx_mod.exe and switch to a new exe launcher for the mod that does all the changes in-memory. So we would not need any patched terranx_mod.exe files anymore. After I deploy the changes, it would be good to gather some insight to make sure it works on all Windows versions. It should be available soon.

[Michael2589]

Each faction definition contains TWO lists of base names. One is a list of generic names, and one is a list of names that are specifically for sea bases. (Have a look at a faction file if this isn't making sense.) The way the vanilla base naming algorithm works, is that:

if the base is a land base, it grabs a name from the generic name list, and if it's run out of those it names the base "Alpha Sector" or whatever. It ignores names from the sea base name list because most of them would look pretty strange on a land base.

If the base is a sea base, it looks for a name from the sea base name list, and if it's run out of those it looks for a name from the generic list, and if it's run out of those, too, it names the base "Alpha Sector" or whatever.

Your algorithm is doing the same thing as the vanilla algorithm for land bases, but for sea bases it's just checking the list of sea base names and, if there aren't any of those left, it gives the base an "Alpha Alpha Sector" style name, even if there are still unused names from the generic list. I know this because I've founded a sea base and gotten an "Alpha Alpha Sector" style name, and then founded a land base and gotten one of the names from the faction's generic list.

Re: Supply pods: That's what I was thinking about monoliths.

Re: directed research and blind research
I certainly wouldn't use blind research in a competitive multiplayer game, as better luck on randomly getting key techs early would almost certainly be more important in determining the winner than huge differences in player skill. However, I use it for single player games specifically because I like the gameplay being randomized.

The reference to Progenitors is because of something that I consider a vanilla bug rather than a balancing issue. Think back to why you created the "tech_balance" option. I don't want to use it in blind research games, for all the reasons that I'm playing with blind research, but I'd certainly use it in a directed research game. But if I'm playing a blind research game with Progenitors, then some factions are using blind research and others are using directed research, and I can't set tech_balance=0 for some of them and tech_balance=1 for the Progenitors. It's a minor and obscure issue though, and it's a problem that exists in the vanilla game, and I could fix it by tweaking alphax.txt if it was that important.

[induktio]

Starting from Dec 10 develop build, Thinker now contains a new launcher program thinker.exe which is used to start the mod.

• This requires the original GOG/non-GOG game binary terranx.exe in the same folder.
• Any patches are NOT written to terranx.exe, instead they are applied dynamically at runtime. Players can use terranx.exe to start the unmodified version of the game if necessary.
• Command line arguments should work as before. They are now given to thinker.exe.
• It is possible to use terranx.exe that is pre-patched with Scient's Patch v2.0 but any other versions of custom patches will probably fail at startup. Note that this is not necessary because the mod will apply Scient's Patch v2.0 changes at startup anyway.

If you have any issues with the launcher program, let me know. It could use more testing on different Windows versions, but probably it will be okay.

About the base naming scheme, I can see there's a difference with the way it works on sea bases vs. vanilla. However I can't decide now how to rewrite the code, because it uses a specific randomization scheme which should probably maintain compatibility with vanilla version. Have to think about it later.

[Michael2589]

I completely forgot to mention my experiences with the launcher program.

I'm running on Windows 10 Home version 21H1, Bitdefender antivirus, Intel Core i7. The first time I tried to start it, Windows recommended that I not run it because it's not a recognized program, and it was flaky the first few times I tried to start it (sometimes it would start, sometimes the program would just close before it did anything for no apparent reason), but since about the fifth time I started it, it's been fine. The issues may be because I was setting it up in a new folder to continue playing a game from an old folder, so Alpha Centauri.Ini was pointing at a game in a different folder, etc.

It actually works better than the previous version, because BitDefender was absolutely convinced that the older thinker.dll's were malicious software, but it's fine with the thinker.dll that comes with the launcher program.

[induktio]

Good to know it works well with Windows 10. I did not test that Windows version myself so that's important. I'm mostly finalizing some changes so we can bundle these in a new release.

There's been some issues with antivirus programs flagging the mod files with false positives. Antivirus heuristics become really nervous when they see some program patching binaries in memory, usually the detection is like HEUR/Malware or something generic. Not sure if there's a way to get completely rid of them other than maybe providing signed binaries, but that's a hassle. In any case there should be only very few false positives for thinker.exe now.

[induktio]

Version 2.9 is released now, but there's still some areas in the game balance that could be adjusted for next release. These could also involve additional alphax.txt edits.

Another player commented on reddit it's (too) easy to rush projects with crawlers. The AI doesn't currently rush projects with energy reserves or use crawlers for hurrying them either. This could be adjusted so that it would be harder to just grab every project later in the game, but it needs to also be balanced against how fun such an AI would be to play against. Thinker is not entirely about min-maxing for victory even though most of the mod is such.

• Full mineral disband ability could be removed from crawlers and make them disband the same amount than other units. When crawler upgrading is possible with energy reserves, this would be just too strong otherwise. Their purpose should be to gather resources, not be instantly disbanded after upgrade tricks.
• AI should not need to use crawlers to rush projects in such case, energy rushing would be sufficient.
• AI could sometimes use their probe teams to subvert (lone) enemy units, but the implementation side binary stuff is not entirely reversed.
• AI currently doesn't evict neutral (non-hostile) probe teams from their territory. Maybe this should be implemented the same way the player can evict foreign probe teams from their territory?
• Config option limit_project_start could be removed. Note it doesn't affect the top 3 difficulty levels at default setting, but the option is kinda redundant for this kind of mod.

[Michael2589]

It's not exactly stack damage. If, for example, the AI has 3 military units and 2 probe teams inside a base, and you defeat the first military unit, there's no damage to the other units. Defeat the second military unit, same thing. Defeat the third, and the probe teams die at as well, and if you're looking in the right place on the screen you'll see the message "3 units were destroyed" briefly displayed.

I know it's weird. (That's why I mentioned it.) But I've been paying attention to it, because it's the difference between being able to use armoured probe teams as maintenance-free base garrisons or not.

I can't remember now whether this is how it works in vanilla.

[induktio]

Hmm. Thinker does not change those stack damage rules currently. I think that probe team stack destruction rule should be removed at least when they are inside bases. It doesn't make much sense. Seems almost like an oversight in the game. If I remember right, it's processed somewhere in the long and convoluted battle_fight function. Outside the bases it could maybe apply some rule that if an armored probe is defending the tile, stack damage destruction does not apply or something. But I'm not sure how easily those can be patched in.

[Michael2589]

I think it's completely reasonable as a feature. That is, being able to use maintenance-free units interchangeably with military units that cost maintenance as defensive troops has a lot of implications. Especially since the armored probe team is actually an offensive threat, since it can subvert units, and can move while ignoring zones of control, and all of that despite being cheap... they don't need to be made stronger.

[induktio]

On the defensive side, that feature shows probes are not that strong. Supposedly if a base contains 5 armored probes and one garrison, they all get wiped out after the garrison loses which seems nonsensical. Outside bases (or bunkers), you could make that argument they need not be made stronger, but that's a stretch to say the current version is a good mechanic for base tiles. At one point I thought about implementing this earlier, but decided to skip it back then when coincidentally I implemented some mechanics that force the AI to disperse its troops into multiple tiles. That doesn't help with stack destruction on base tiles though. Might be a good idea to have some kind of a config option for that.

[induktio]

On another note, just looked throught the base production code and confirmed the game engine has extra AI cheats for facility maintenance. The AI pays only 1/3 maintenance cost for facilities on transcend and 2/3 on thinker level. On other levels the costs are same as for player factions. The numbers that the game shows on infiltration screens do NOT account for AI bonuses on maintenance, the bonus is just quietly calculated during turn processing. These are probably not documented at all, at least I don't remember seeing them mentioned. Seeing how the AI accumulates credits on transcend level can look strange. The bonus on that level is a little high and might need some rebalancing.

[Michael2589]

Obviously, they're not that strong as base defenders (i.e. probes can't be used as base defenders against a military assault). But they shouldn't be useful against military assault. That's not their purpose.

Why should I build base defenders that cost maintenance if I can build base defenders that don't?

As for why it makes sense, the fiction is that they are probe TEAMS, not probe regiments or probe divisions. The number of people in them is considered trivial compared to the number of people in the military units.

It would be reasonable for them to escape a falling base (e.g. use the function that sends them home after completing a mission to send them to a different base owned by their faction, though what if this was the faction's last base but the faction is still alive due to colony pods?) It is not reasonable for them to prevent enemy military units from occupying an otherwise undefended base.

Something that would be nice would be to increase the cost of mind controlling a base depending on the number of military units in it. At the moment, if you have two military units in a square you can't subvert them, but if you have 20 military units in a size 2 base you can get the base and the military units for the cost of mind controlling an empty size 2 base.

I remember the 1/3, 2/3 maintenance on AI base facilities, although I can't find any mention of it. Supposedly it was bugged in SMAC, so that the AI paid triple maintenance on their base facilities at transcend difficulty in SMAC, and this was fixed in SMAX.

I don't think there's a need to change this. It makes sense in that the AI isn't checking that it has enough revenue to pay for its maintenance, and on higher difficulty levels it'll randomly build more of everything, so more base facilities, so more maintenance.

Can you check whether there is actually a research penalty associated with punishment spheres? As far as I can tell, looking at the various screens in game, that isn't implemented, but it's entirely possible that it is and it's just not being reported on the report screens.

[induktio]

Not all of the effects are shown in the GUI obviously, if we dig into base_research this piece of code is run there:

v1 = current_base_id;
v2 = current_base_ptr->labs_total;
bitmask(FAC_PUNISHMENT_SPHERE, &v8, &v7);
if ( (unsigned __int8)v7 & bases.facilities_built[308 * v1 + v8] ) {
v2 /= 2;
}

Some people would like to change the mind control mechanics, but opinions differ, so go figure. Actually when you noted about the probe ratings and mind control, I implemented a small priority change that keeps track when the last base was lost to mind control. If it was recently, social_ai should put additional weight for probe rating.

This was actually really simple to implement because there was already code related to it. This and some other tweaks for loan diplomacy dialog are implemented in Dec 22 build. The AI has new dialog now when they strongly reject a proposed loan.

[Michael2589]

So the research penalty for punishment spheres was implemented, but, whether or not it is buggy, I have no idea. I'm visiting family right now but when I get home I'll look into setting up a test for it.

Thanks for the change to SE if they lose bases to probe teams :).

If that's the same terrible loan offer that the AI offers the player, good riddance. Now if you could only stop the AI from asking the player for loans that would be perfect.

I was trying to figure out situations in which a loan could be a reasonable idea. In faction.txt, one of the abilities you can give to a faction is interest on their energy balance. So if a faction had the ability to collect, for example, 10% interest on their energy balance, it would make sense for them to suggest, to factions they don't dislike, that the other faction give them a loan for (in that example) 5% interest, and then the principle would get repaid as a block after, for example, 10 turns. The faction acting as banker would then keep the excess interest. But since no factions actually have this ability the situation won't arise. And I can't think of any other situations where giving loans actually makes sense.

While you're adding features to the sunspots event to flesh it out, you could give a configuration option in thinker.ini that sunspots would also cut commerce in half. If things are so bad that faction leaders can't communicate, the presumably commerce would be adversely affected as well. I do like the idea of leaders using the cover of sunspots to nerve staple troublesome bases!

[induktio]

Where do you get the idea sunspots work somehow differently in the mod? It should work similarly to vanilla game, you can nerve staple bases without commerce penalty when sunspots are active. Also that DONEPROJECT has a similar sunspots check before it is shown.

So right now the AI affinity for loan granting depends heavily on treaty status, diplomatic friction, integrity blemishes and the relative size of the factions (they are more willing to loan to smaller factions) plus a couple of smaller factors. The diplomacy code is quite convoluted and mixed in with the GUI code, so I'm not going to attempt to replace large fractions of it, but some smaller tweaks on some specific areas could be done.

[Michael2589]

The reference to nerve stapling bases without commerce penalty when sunspots are active was a reference to the change in AI behaviour in patch 2.9. It wasn't intended to imply that you'd changed the game rules.

[induktio]

Previously Thinker Mod has not modified the tech tree but I've been developing some interesting concepts recently. The files can be downloaded here. There were some very specific design ideas that I decided to implement. I have not decided if I should make this new tree the default in the mod, probably needs more comments from players.

• Provide some in-between option for smacx and smac-in-smacx mod with the amount of game content.
• Rearrange all weapon and armor techs such that they are discovered in a linearized sequence, one for each type, no exceptions. Each type has a multiple tech delay before a better one can be discovered.
• Remove all expansion weapons/armors. They just clutter the tree with redundant options.
• Remove all expansion techs (and by addition, most of all references to Progenitors).
• Keep all expansion abilities/facilities/secret projects but move them to appropriate vanilla smac techs.
• Add Global Energy Theory and Inertial Damping techs that were cut from the original game. These are good techs for some of the smacx items.
• As it can be seen from the graph, a special property of this new tech tree is that techs from level n have only prereqs from level n-1. First 7 levels also have exactly 7 techs each.
• This approach forces to prioritize early-to-midgame techs. For example Neural Grafting arrives later than in the base game but I think it's okay. For early game one has to now choose carefully which special ability to put on the units.
• Hovertanks and gravships arrive slightly earlier than in the original game. Graviton Theory could even be earlier but not really sure.
• Some important reorderings: satellites are now discovered in a sequence Fusion power -> Orbital Spaceflight -> Super Tensile Solids -> Advanced Spaceflight. Biomachinery is moved later because it contains an OP project, and so forth.

[JoGr223]

"Rearrange all weapon and armor techs" I don't think it solves any actual problem in the game ?

"Remove all expansion techs" I don't like it, it looks like solution in search of a problem; I like that for example PEG SP has its own tech (Adaptive Economics) as well as it gives thermocline transducer

"Keep all expansion abilities/facilities/secret projects but move them to appropriate vanilla smac techs." that's the key thing and you are not showing how you are going to do this

"Hovertanks and gravships arrive slightly earlier than in the original game. Graviton Theory could even be earlier but not really sure." you are not stating what's the point of all these changes; the hovertank is actually what i've been thinkig about too (and i think one of the mods might already have that); but the problem here is with Doctrine Air Power - it would be better if hovertanks were before planes

I think you are making deep changes without any reason - each of this proposal should have separate topic :) it feels rushed and too global; better to make small directed changes where each solves something; I originally thought Thinker mod is only about AI, then I learnt it makes slight rebalance like slower research and no attack bonus from better reactors; I'd prefer to make only necessary changes; readjusting whole tech tree globally does not look like something that's needed

[induktio]

Let's just clarify a couple of things. You can click on the dropbox link and it should contain the exact same alphax.txt that was used to draw the graph. You can also install it in the game and look at datalinks, so it's not like the implementation is "hidden" in any way. Planetary Energy Grid would get moved to Global Energy Theory and it's quite a fitting tech for the project.

Also let's be clear about the fact that the mod will not require having any specific tech tree while playing. Already the mod ships with smac-in-smacx as an alternative, so the new tech tree might get added as an additional choice. If some people like to have just the default settings, there's nothing preventing it. Of course then there are the binary patches, but thinker.ini file is for them.

There are many players who may not particularly like the expansion stuff but would like to keep some of them. And some of them just play smac-in-smacx. Even though I didn't describe all of the rationale there, I don't really see how you can say they're just unnecessary changes. :) For example removing the expansion armors/weapons has already been done in Will To Power, because those items basically duplicate existing stuff and add very little to the game.

Also weapon/armor reordering serves a specific purpose: it makes each of them have a specific timeframe where they can be used. In the vanilla game some of them nearly always get skipped over because a better one is available at the same time. Sure, there are many changes, but also on other hand, I've tried to keep the previous ordering for many techs where possible. So here we can discuss it. :) That hovertank/needlejet reordering could be possible I suppose, that's something to think about.