SMAX - The Will to Power - mod

[Alpha Centauri Bear]

Has some alteration been made that's bugged the Spartan's ability?  There are a few thinker.ini variables affecting prototypes, but they don't obviously govern this.

Thank you for discovering this. It was a pretty hard coded ASM stuff prone to mistake as advertised. 😄
https://github.com/tnevolin/thinker-doer/blob/master/wtp_changelog.md#version-183

[bvanevery]

New game in progress, begun with version 182.  Huge map, 30%..50% land, average settings, random opponents.  I drew the Cult of Planet.
expansion_factor=25
faction_placement=0
alternative_combat_mechanics_loss_divider=1.0
combat_bonus_territory=15

Details bountiful isle

bountiful isle

I was given an island.  It had some fungus on it, but was also lushly green and has Mt. Planet.  Can't complain.  That's 3 for 3 with the stock faction placement algorithm giving good land.  This supports my theory that it evaluates the quality of the land given, and the alternate algorithm does not.

MY 2190.  I am at war with the Caretakers immediately to the south of me, and allied with the Data Angels far away from me.  I'm doing the usual Fundamentalist thing as it's clearly a slam dunk as long as you have an ally.  It's also easy to get an ally when you're pathetic, as some faction is likely to see you as a sad sack in need of help.  This is almost exploit territory, but I don't feel that bad considering how imbalanced and painful most of the SE choices are.

Due to Completions I have a functioning Skunkworks, and I completed some real armed ships.  I thought I was going to cream the Caretakers, but surprisingly, they fought back with real ships.  The initial clash is a bit of a stalemate, and I'm about to probe.

Most factions can't get access to Non-Lethal Methods units, but I have the Worm Police ability.  It is extremely helpful, almost exploit territory.  I've got a pathetic indigenous life form in most of my cities, adding to the crowd control.

The Data Angels only have 16 cities.  They have 1 active Colony Pod and 1 in production.  1 active Rover Colony Pod and 1 in production.  They are mostly building Trance Scouts and Recycling Tanks.  They don't appear to have any land neighbors, nor are they at war with anybody.  Morgan is visible in the sea to the south, and Roze sold me his comm frequency just before allying with me.

Morgan is in contact with the 3 other factions, but wants an exorbitant 125 credits for the commlinks.  Later for that!

Details my buildership

my buildership

I've managed to get some SPs done, using up a bunch of Artifacts.  I will also soon complete the Universal Translator.  That's not worth spending anything on, because cashing 1 Artifact is a cheaper way to get a tech.

Next year, the Caretakers annoyingly sink my foil probe, and someone calls an election.  Deirdre declares war on me.  Miriam and Aki Zeta-5 give me Treaties.  After some trading, I talk to Miriam again and she allies with me.  She likes my godly style. 

The Believers have 23 cities.  1 active Colony Pod.  2 active Rover Colony Pods, 2 in production.  Their empire is mostly Scouts.  They aren't at war with anybody.

[bvanevery]

MY 2199.  I got Morgan to ally with me.  He has 30 cities.  3 active Colony Pods, 4 in production.  1 active Sea Escape Pod.  3 Sea Colony Pods in production.  2 Fusion Sea Colony Pods in production.  I'm starting to think that expansion_factor has no effect.  His empire is almost all Trance Scouts.

I notice that alphax.txt has:

Code: [Select]

Trance Scout Patrol,    Infantry, Gun,          Scout,      2, 0, 0, Brain,   -1, 00000001000000000000000000
Based on my experience with the stock binary, I think it is exceedingly likely that the AI is obsessed with producing this unit.  To the point of producing almost nothing else.

[Tayta Malikai]

Hmm. I see a lot of controversial responses on recent AI changes. Let's review them.

Thinker main goal was to improve AI and WTP continues on this path as well. I have heard a lot of complaints about some OP exploits humans can do but AI cannot. Teaching them to AI is usually the best way to resolve that problem. However, this is also the most difficult option. In absence of which, we usually resort to limiting such exploits to equalize human vs. computer capabilities. ICS was one of the most notorious one everybody tried hard to limit. Yet now when I was able to teach computer do the same, making it no more exclusive human exploit, it doesn't seem to be unconditionally liked. What is going on, guys? Were you hypocritical when you complained about it and demanded this problem to be solved for good? I have spent quite substantial amount of time and mental efforts on it. Was it for nothing?
Let me know if this is the right direction and/or something need to be done with spamming as well, regardless of AI success.

Keep in mind that neither Thinker nor WTP planned to put any limit on AI improvement posing themselves as challenging mod. So the argument that game becomes unbeatable on toughest level, while pretty valid for game attractiveness in general, is not applicable to this mod.

About AI disregarding defense while expanding without limits. That is true. Building more colonies obviously impacts combat units. This allowed me to amass an army and easily cut into neighbor territory. I plan to work on defense more. However, even in current state I can eat some number of surrounding bases and I may even match other factions in number of bases but it takes long and I had to put everything into building combat units so this seriously delays my economical development. By the time I an about to eat next faction they become very strong and easily block my advancement. Tayta reported that she can continuously beat them all conquering whole planet, though. I would love to see her SP game progress just out of curiosity.
Nevertheless, it seems that AI grow stronger comparing to previous settings, overall.

I would like to see my SP game progress, too. Haven't had much time to play lately... and what time I do have goes to playing my turns in MP. And next year I expect most of my non-MP time will be spent trying to mod SMAC features into Terra Invicta, when that comes out. (Making my own mods instead of complaining about other peoples', that's right!).

This was about where I got up to when I stopped playing. The short version is I took two bases, Edenists killed a former, and then offered me all their money in exchange for peace. I took the offer because, well, why build military when there are facilities and projects to be built? Those Tree Farms and Genejack Factories and Research Hospitals aren't going to build themselves.


My assessment is that I could definitely win a conquest victory like this, it would just take a long time and probably become tedious with how many sea bases the AI has. I'd probably settle for taking over their mainlands, then sitting back and building up for the diplomatic victory. This seems to be the outcome you're angling for, but do correct me if I'm wrong.

It's definitely worth repeating that the AI performs a helluva lot better now when it comes to economy and research. The AIs that were able to expand with a gazillion sea bases and enjoy mega commerce with each other are raking in the cash and tech big time. So great job on that.

It may be that the AI performs so well economically now that the player is forced to devote all their resources to keeping up, making war less feasible. That sounds like it was your experience too.

I'm not super opinionated on where the core mod ends up going, since I'd most likely just end up using a modified version for "my" solo games anyway. I do of course appreciate all the new features and configuration options you keep adding. Things like infiltration expiring make it much harder to get complacent now.

[Alpha Centauri Bear]

Not angling any specifics. Just teaching AI whatever tricks human knows to make it at least look smarter in a name of Thinker.
Stronger empire means more of everything. Not only energy and research but proportionally more units too. They just seem to sacrifice some current military power for bigger future one. That may be perceived as temporary disadvantage. I am going to work on this.

[Alpha Centauri Bear]

https://www.reddit.com/r/4Xgaming/comments/jo028q/alpha_centauri_bombarding_empty_square_with/

Load the save and directly attack the tile with two enemy cruisers, 12-1-6 and 1-12-6. Use a gatling 5-1-6 cruiser for this. Instead of picking the unit with better armor and easily fending off the attack, the game will pick the unarmored one to defend, which will lose immediately.

Also. Pick a free tile in a 2-tile range of a enemy cruiser or artillery. Bombard it. See what happens on successful bombardment.

(a -100% hasty penalty, resulting in the immediate death of your ship, as it expended the moves and can't defend. It is not supposed to happen, I think. If there is supposed to be an artillery duel, then let them duel before the moves are expended)

Found some questionable stuff in code. Now trying to understand how it is supposed to work. See my reddit post.

[Nevill]

Oh. Remember the entry "Spore launchers will sometimes attack each other"?

Well, would you look at the screenshot.

My 2-14-1 arty is bombarding a tile near the believer 10-14-1 arty, yet it is my OWN artillery (30-14-1) that ends up intercepting and initiating the duel.
In particularly egregious cases, the artillery may initiate a duel with ITSELF at -100 Hasty penalty, and lose.

I can offer no opinion about the authors' intention, and the wiki article may have been written to try and explain an existing fact. There are no references to duels over the tiles that contain no units, and tiles with units never get defended in that way even though there is no logical explanation for why arty units should protect empty tiles and not their allies.

...I like the idea, though. But I think the entire concept is born out of a bug.

(Then again, I consider 1-round land/sea duels a result of a bug as well. The wiki reflects what is in the code, not what should have been there.)

[Alpha Centauri Bear]

(Then again, I consider 1-round land/sea duels a result of a bug as well. The wiki reflects what is in the code, not what should have been there.)

Or it could be the opposite. At least looking at the code does not give me the clear feeling which branch is a bug.

[bvanevery]

MY 2241.  I completed the Maritime Control Center.  I've mostly developed my island and have only sent completed prototype units to fight the Caretakers.  2 decent ships really terrorizes a bunch of Scout garrisons.  In the past they have occasionally sent a real ship with an armament to try to fight back, but there's been nothing sustained.

Details invincible probe team

invincible probe team

Now they've finally sent this wunderkind probe team unit.  I've already bombarded it with a Fusion Laser.  That didn't make any wound at all, which seems pretty lame to me.  I have 8 offense, they have 6 defense, seems like I could have at least done 5% or 10% damage or something.  But nope, nada, zip.

The odds on this matchup are crazy bad.  We're almost equal, yet I'm given only a 15% chance of victory.  Since I don't object to the Territory bonus, I say the Alien Defense bonus is too high.  In my mod, I removed both the Alien Offense and Alien Defense bonuses.  I didn't want grossly superior aliens like in the stock game.  If you want to keep those bonuses, I'd suggest knocking them down to 15%, equivalent to a PLANET bonus.  They're pretty much magic excuses and it's dumb.

What bugs me isn't that the Caretakers will liberate their bases with a probe.  I'm surprised they didn't attempt that a long time ago.

What bugs me is that they never pay the real cost of captured military units when they do so.  These are expensive ships with good armaments on them, and it should be expensive to take them over.  Otherwise the whole game becomes about probe team exploits and that's boring.  I have rage quit SMAC over the years more times because of this problem than any other.  In my mod I doubled the cost of all mind control, which somewhat helps, but it's not an ideal fix.  The real problem is the complete lack of accounting of units in and around the base.

I decided to retreat to Harmonic Acolyte and hope I win an artillery duel against my own unit when it is captured.  Suiciding against the probe team doesn't seem likely to stop one of my bases from being captured, as even 2 movement is enough to do the job.

Details spared

spared

MY 2242.  I'm surprised my Fusion Megafoil was not swept up in the probe.  In the stock binary it would be toast.  If WTP has improved the behavior for this, I'll take it!  I think that expensive probe ship must have died too, because I don't see it at any nearby sea base.

[Alpha Centauri Bear]

https://github.com/tnevolin/thinker-doer/blob/master/wtp_changelog.md#version-185

Closing this question for good. I have discovered that destroying improvements (bombarding empty square) explicitly exhausts movement points. Therefore, it seems that it was not a design decision to protect empty squares with enemy artillery.

[Nevill]

Or it could be the opposite. At least looking at the code does not give me the clear feeling which branch is a bug.

Can you tell us what you see? In layman's terms, so we could at least try to understand the intent.

I'd like to know what does choosing the least defended unit (the message you quoted in the post with the reddit link) have to do with it.
Edit: Forget it. I can't read. 

Right now I see several inconsistencies:
- The artillery duels are initiated over empty tiles only, ones that are "protected" by another artillery. Tiles with units are not protected.
- The duels do not care for the faction of the artillery (and sometimes they don't care if there even is another unit)
- There is no actual duel. AFTER the bombardmebt succeds, the movement point is simply expended and the attacking artillery is subsequently killed as its power is reduced to 0 by the -100% Hasty penalty.
- A bombardment has to succeed for the duel to commence. If the bombardment fails, the duel is not initiated. How do you fail bombarding an empty tile - the success and failure look exactly the same?

What would be a logical example of code?
1) The attacking and defending units could duel only if each is within range of the other, i.e. no more than 2 tiles apart.
2) The duel would only commence if the attacker chooses to bombard a tile within the range of the defender (different from 1, as though the units could be within a 2-tile range of each other, the attacker can choose to fire in the other direction). The battle would play out the same as if the attacker chose to fire directly upon the defender.
3) What is in the tile should have no bearing on whether the duel commences. Neither should the success of the bombardment. If you choose to fire upon a tile within the sphere of influence of another arty, and you yourself are within range to fight, you fight.
4) If there are several artys eligible to protect a tile from bombardment, the one with the highest chances gets picked.

That would turn arty units into some kind of land-based interceptors, able to protect an area in their vicinity from enemy bombardments. Even then their usefulness would be limited, as the area would most likely be bombarded from outside their very, very limited range.

I just can't envision how this mechanic would work or what the point would be.

Edit: NVM, ninja'd.

[Alpha Centauri Bear]

The odds on this matchup are crazy bad.  We're almost equal, yet I'm given only a 15% chance of victory.

You are not almost equal. They are 50+% stronger.
That is what you get with vanilla combat mechanics. A slightly stronger unit is almost a sure winner. Have you experienced any different in decades of playing??? Why suddenly start complaining about it in the mod play testing that has nothing to do with this mechanics?

Higher randomization in WTP serves the specific purpose of giving more winning chance to weaker unit.

[Nevill]

What's up with these ability costs?  There's no predictable consistency between abilities that cost 16 and abilities that cost 1.

It confused me at first as well. You may have figured it out on your own, but just in case - cost factors 16/32 denote the cost in additional rows, not an added percentage.

Cost x = Base cost + 25% * x
Cost 16*y = Base cost + y mineral rows

Base cost 60 would give cost 80 and 70 for cost factors 1 and 16, respectively.

It's probably done that way because it's hardcoded somewhere.

[Alpha Centauri Bear]

Can you tell us what you see? In layman's terms, so we could at least try to understand the intent.

There is a "destroy" function that, apparently, only get called when bombarding empty tile.
There are battle/fight function chain those are called for unit-to-unit combat in all other cases.

The destroy function exhausts unit movement.
After it there is a condition check on global variable that falls into normal unit-unit combat if variable is set.
Variable is NOT set when you start the program.
Variable gets set in different combat scenarios like direct attack.
It never gets cleared! Meaning it corrupts the whole program in memory. Reloading doesn't fix it!

Tiles with units are not protected.

Yep. That is a big mess that directly contradicts the description.

What would be a logical example of code?
1) The attacking and defending units could duel only if each is within range of the other, i.e. no more than 2 tiles apart.
2) The duel would only commence if the attacker chooses to bombard a tile within the range of the defender (different from 1, as though the units could be within a 2-tile range of each other, the attacker can choose to fire in the other direction). The battle would play out the same as if the attacker chose to fire directly upon the defender.
3) What is in the tile should have no bearing on whether the duel commences. Neither should the success of the bombardment. If you choose to fire upon a tile within the sphere of influence of another arty, and you yourself are within range to fight, you fight.
4) If there are several artys eligible to protect a tile from bombardment, the one with the highest chances gets picked.

Yes, I would like the same to work. However, it is sooo messy as you can see already, I don't want even study this code much less touch it. For now, I think the fix for bombarding empty tile is sufficient enough. We can forget about whole artillery interception functionality then.

[Alpha Centauri Bear]

What's up with these ability costs?  There's no predictable consistency between abilities that cost 16 and abilities that cost 1.

It confused me at first as well. You may have figured it out on your own, but just in case - cost factors 16/32 denote the cost in additional rows, not an added percentage.

Cost x = Base cost + 25% * x
Cost 16*y = Base cost + y mineral rows

Base cost 60 would give cost 80 and 70 for cost factors 1 and 16, respectively.

It's probably done that way because it's hardcoded somewhere.

Yes. There is only one field for ability cost and I split it bitwise on two values. This is described somewhere in readme, I believe. I also plan to display more human understandable text in designer as well.