Fission Armor mod (SMAX)

[bvanevery]

With weapon-armor equalizing I cannot even imagine how much attack advantage you need to achieve that on top of Airspace Complex and AAA tracking which already quadruples defense.

AIs in the stock binary do not achieve that with regularity everywhere.  Many cities have no AC.  You will see AAA units but not necessarily piles up on piles of them.  When a city is stacked full of units, it's usually not full of AAA units but something else.

I'm not sure if it's possible to get very far ahead of the AIs with a weapons advantage in my mod.  I've made changes, but I've only played so many games in the past few months having made those changes.  When I write something up, you may at times hear me say, "I don't even know how to play my own mod."  I really don't.  I'm not sure what's optimal or not.  I don't quite know what I can exploit or not.  Like the last test game I quit, before kicking 1.27 out the door today, both Morgan and the University were doing really well.  They were distant and I wasn't managing to get anywhere near them.  I had started to wander a rail out across a large land mass, trying to feel a way to the enemy, but I didn't even have a route completely scouted.

I meant potentially deep resources. When one faction is very determined to throw everything at enemy they overcome the defense anyway.

When an AI tries to do that to a human, it often doesn't work.  They will move in a big stack, and the human knows how to exploit that, wiping out vast quantities of AI units in 1 stroke.  I've had some truly horrific waves of enemies thrown at me in the past, that I just weathered and shot a lot of 4-1-1 and 5-1-1 units at, until the AI wore itself out.  Then finally moved up the road for the kill, after many years of waiting. 

The only kind of wave I haven't been able to stand up to yet, is waves and waves of mindworms in a flood.  It's beyond AI production, it doesn't obey any obvious production rules, it's just complete abuse.

Can the AI stand up to another AI onslaught?  At some point, no it can't.  I'm not sure what point that is in my mod anymore though.

[Alpha Centauri Bear]

With weapon-armor equalizing I cannot even imagine how much attack advantage you need to achieve that on top of Airspace Complex and AAA tracking which already quadruples defense.

AIs in the stock binary do not achieve that with regularity everywhere.  Many cities have no AC.  You will see AAA units but not necessarily piles up on piles of them.  When a city is stacked full of units, it's usually not full of AAA units but something else.

That is what I said elsewhere. Extreme combat effectiveness depends on opponent unprepared defense, not on your technical advantage. If their defense is weak then you don't need choppers to win. They just may do it a little bit faster.

I meant potentially deep resources. When one faction is very determined to throw everything at enemy they overcome the defense anyway.

When an AI tries to do that to a human, it often doesn't work.  They will move in a big stack, and the human knows how to exploit that, wiping out vast quantities of AI units in 1 stroke.  I've had some truly horrific waves of enemies thrown at me in the past, that I just weathered and shot a lot of 4-1-1 and 5-1-1 units at, until the AI wore itself out.  Then finally moved up the road for the kill, after many years of waiting. 

That's right. Human defend actively. That alone like triples the defense effectiveness. With fully stacked bases (artillery, counterattackers, ECM, AAA) you don't even need walls sometimes. I was comparing passive defensive bonuses with row attack potential. Sort of AI vs. AI. Human tactics is a different story.

[bvanevery]

I will keep my eye peeled for whether I think "Choppers would be far more effective here".  But at this point in my mod, I've tried to push Gravships more into the realm of the usable, and I've made them faster.  Foils speed 5, Cruisers 8, Locusts (on Chopper platform) 8, Needlejets 10, Gravships 12.  I am unlikely to go back unless I hear playtester howling.  Originally I was going to try much longer range aircraft, but I decided to kick it out the door more conservatively, so that I wouldn't have to belabor the testing.  I just couldn't get to it.

[Alpha Centauri Bear]

Played my mod a little bit more and realized my way of pricing SP is flawed. My initial driver was to not let AI build SP too fast (almost instantaneous) so that human player has opportunity to take actions if they like to. That should be a baseline for project pricing. The actual value should add on top of that and modify it a little bit. So that all projects can be built by AI in 20-40 turns with their current best city production power. Anything less than 20 turns gives little flexibility to human player. Whereas anything longer than 40 turns is ridiculously long project (in my opinion). I am not sure AI use crawlers effectively to speeds projects up, though.

Here how complete picture should look like in my vision:
Different era SPs should take roughly same build time in their corresponding era.
Same era SPs should differentiate slightly by their usefulness. This will also result in cheaper projects to be built first.

[bvanevery]

I did all that / tuned all that.  Happy with the results.  In my regime of expensive SPs, you really have to choose which of the big ones you're going to go for.  At least by midgame, the AIs are going to get the others.  If you survive and prosper in the late game you'll surely be the winner IMO, so everything becomes more of an honorarium / sandbox.  However I am biased by being too good at the game, and not representative of all players who will try my mod.  So maybe they'll still have difficulties late game for all I know.  I haven't had enough playtesters verbalize their experiences to know for sure, but one recently did seem to be getting challenged for a longer period of time, when I would have summarily won the game by then.

[Alpha Centauri Bear]

I did all that / tuned all that.  Happy with the results.  In my regime of expensive SPs, you really have to choose which of the big ones you're going to go for.  At least by midgame, the AIs are going to get the others.  If you survive and prosper in the late game you'll surely be the winner IMO, so everything becomes more of an honorarium / sandbox.  However I am biased by being too good at the game, and not representative of all players who will try my mod.  So maybe they'll still have difficulties late game for all I know.  I haven't had enough playtesters verbalize their experiences to know for sure, but one recently did seem to be getting challenged for a longer period of time, when I would have summarily won the game by then.

Thanks for reply. Nice to know. I'll examine your SP costs and see if I can use them as insight. Do you have any rationales about SP appearance time as well? Would like to synchronize on this too. Maybe in you readme or version changes description?

[Alpha Centauri Bear]

expensive SPs

I looked at SP prices in your 1.36 version.
Wow. When you said your SP are expensive I thought beef them up a little but found the opposite. Actually most of my projects are more expensive than yours sometimes by far. This is actually one thing I'd like to review and make sure they are not insanely expensive. However, in the mid and late game the combination of base row production powers and crawlers makes it piece of cake regardless of the price.

I'll review them in light of following philosophy.
SPs were designed to be lucrative and race for them is one of the essential elements of the game. I don't think their price need to outweigh their benefit exactly. It would be senseless effort anyway. Rather it may reflect their relative usefulness within the era to rank them up from easy to get and less beneficial to hard to get and more beneficial. Whereas between eras their base average cost should grow to reflect increasing producing power to keep their build time roughly unchanged. As I pointed out earlier, it should be somewhere in range 20-40 turns to represent both relative toughness of the SP comparing to regular facilities and closed enough time period. With more than 40 turn building time it slides to next era essentially delaying its benefits.

In this philosophy fast research and early acquiring SP technology is still crucial for SP dominance as in original game. I guess this is nice to keep it like that to maintain certain value for both raw production power and research speed. Faction with fast research got to chose which project to go for. Whereas faction with production power can collect rest of project in larger quantities or even overtake earlier starter with higher production speed if they have a chance to.

[bvanevery]

Bear in mind, in my mod you aren't going to get supply crawlers, factories, or thermal boreholes until late midgame.  You're going to have to do it with 1 city only.  You are not going to get any help chucking out my 500 mineral SPs like the Planetary Energy Grid and the Hunter Seeker Algorithm.  You are probably going to get one and not the other, because you won't have the tech or the money to do both.  The main way to get SPs done in my mod is to build up a lot of money.  Often assisted by burning through Artifacts to get things done before the AIs do.  This makes Democratic, Capitalism, and Wealth rather valuable.  One tester complained that my versions of Capitalism and Wealth are rather dry and unexciting, but my counterargument is they simply haven't played my mod enough, and don't know how important it is to buff ECONOMY to get more money to finish SPs.

All of my SPs are occurring at the point in the game I want them to.  I'm very satisfied with that and have done many iterations on it.  I also use SPs to obstruct the AI from trading key techs with the human player.  It usually works, although every once in awhile the AI gets a hole in its head about the value of some SP tech and trades it away.  That may be happening early game and may be based on a mistaken belief about what the AI thinks it can get built.

A big flaw of the stock AI is it doesn't know how to rush SPs with money, even when the AI is fabulously wealthy.  The AI only seems to know how to do it with supply crawlers.

[Alpha Centauri Bear]

True.

[Alpha Centauri Bear]

To the all readers of this post.
I am thinking about modifying the in game unit cost calculation to make it more reasonable and give text modders more flexible options. To that end I am collecting feedback from interested parties.

Usable combat unit types I consider balancing

• Scouts: infantry, mobile, sea.
• Infantry base defender.
• Infantry field unit protector. Same as above but probably extremely beefed up against air and mobile attacks in a field.
• Infantry pure attacker with minimal defense.
• Infantry protected attacker. A hybrid of two types above.
• Infantry artillery with minimal defense.
• Mobile versions of all the above except base defenders.
• Sea versions of all the above including base defenders.
• Needlejet pure attacker. Needlejet defense is mostly useless: Air-to-air combat is resolved by comparing attack factors—the armor value is ignored.
• Chopper pure attacker.
• Chopper protected attacker. It does make sense for chopper as it is vulnerable to ground unit attacks.
• Same for gravships.

I consider all field infantry unit to be available in mobile version for the simple reasoning: if they are all useful then one may want to give them mobility to move them faster provided cost increase is reasonable.

Here are my thoughts so far about unit cost and attack/defense ratio (in incoherent fashion)

The ultimate goal is to give economical advantage to defender. Conquering same technologically developed opponent bases should cost triple to invader in economical loss. At the same time there blitzkrieg should not be possible by the sheer outnumbering. Both sides should lose units and build/supply reinforcement to the front line to sustain invasion or resistance. That guarantees no player can crush others in short term without actual incurring proportional loss. That ensures victim has time to react and consolidate the defense as well as get help from international society.
All the above does not preclude anyone who is stronger than all other factions combined to wipe out others. However, you need to get to such level of economical dominance using political maneuvering first. Until that any adventurous war can significantly slow down your future economical development.

Same level weapon and armor rating should be about the same. That gives slight advantage for defender in both field and unprotected base.

Unprotected base defense bonus should be at least comparable to terrain one (50%).

Sensor bonus should be high enough to make attacker think about destroying sensors as a preparation for attack but not extremely high to completely prevent base from conquering even by 6:1 unit number advantage. It seems that 50% is a sweet spot. Unfortunately, it does not work on sea. So sea battles should be balanced somehow differently. Other option is to change it back to 25% so land and sea battles does not differ that much. Or even to 0%? I'll take your advice.

Two level more advanced weapon should be no more than twice stronger than armor resulting in lowering loss ration down to 1:1 something. Further weapon advancement may completely eliminate attacker losses. I feel like that is in line with the game idea. Indeed, if you are 4 levels ahead you should be able to beat the crap out of anybody.

Ideally, all war breaking attacking technologies should be preceded by corresponding counters in tech tree.
AAA tracking should appears before any air unit.
Anti-native counters should appear before natives production.
Comm Jammer is fine to come somewhat later than mobile units. As Mobility appears quite early but requires some weapon advancement to become a threat.

Mobile units have two advantages: 1) quicker to deploy to distant location, 2) more likely to be in attacking position and, therefore, are designed as primary attacker. Former gives you strategical advantage in reacting to threats in far apart locations as well as saving you some on paying for units in transit to front line. Latter allows to save on armor making them cheaper comparing to fully armored unit. All together this calls to make mobile unit proportionally more expensive as speed is just a multiplier for other abilities. I'd say somewhere 50-100% more expensive.
Same is true for hovertanks or other movement modification. Higher speed should come with about proportional cost.

I don't see much difference in sea units progression. They work same way on water as land units on land. They should have about same cost model too. Their sheer speed calls for comparing them to land mobile units rather than infantry, though. My rule of thumb is to synchronize fully armored foil to pure attacker speeder and same for cruiser <-> hovertank. I always consider fully armored sea units as a baseline for comparison just because they play as both attacker and defender. Whereas, on land these roles are primarily delegated to infantry for defense and mobile for offense. Of course, one is able to construct cheaper pure sea attacker or defender as needed.

I am unsure about needlejet cost. Their probably should be prices about the same as same strength pure armored infantry with AAA as they are twice as weaker against them. So regular 2:1 attacker to defender ratio. Share your thoughts.

Same thinking about choppers. They are slightly stronger due to multiple attacks but this advantage is nothing when attacking fully protected units and bases. So they are only more advantageous in cleaning out weakly armored units in a field. At the same time they are vulnerable to ground counterattacks. All in all should be no pricier than needlejet.

Gravship should be somewhat pricier than needlejet due to its infinite range, etc. However, at this game stage extra movement is largely irrelevant.

Proposition for formula

Chassis price is a multiplier

Factoring in the weapon and armor values are a little bit trickier. The idea here is that unit uses either one of them at combat encounter, not both together. Therefore, it makes sense to price the unit largely based on their primary (strongest) statistics. Indeed, attacker uses their attack and defender uses their defense abilities most of the time. Their secondary statistics adds about half of its value to unit as it grows. Therefore, fully packed unit should cost about 50% more than its pure versions. Indeed, it seems that mixed unit is more valuable than pure attacker/defender as it can be used as both thus adding tactical advantage. At the same time one mixed unit is clearly weaker than two complimentary stacked ones. attacker counterattacks with full strength after defensive battle regardless of outcome. Whereas, mixed unit loses its attacking potential with health while deflecting the attack. So 1.5 multiplier for mixed unit seems right. Another important consideration to not make this coefficient bigger is to increase AI field units survival rate while they are moving toward the target under air attacks.
Therefore, I came up with the following rule: primary statistics + 0.5 of secondary statistics above minimal.

That is pretty much it for all units. No need for weird rules like halving sea unit armor cost or dividing air unit cost by 4. Chassis type multiplier does it all and exposes this flexibility to text modders.

I am not sure yet how to price abilities. Some of them work effectively proportionally to unit statistics like comm jammer or blink displacer. It makes sense to price such abilities proportionally. Whereas others are completely unrelated to unit base cost like anti-native enhancements those usually placed on weakest units to save on cost while having same effectiveness. Another example are things like algorithmic enhancement which price model is quite unclear at all.

[bvanevery]

> Conquering same technologically developed opponent bases should cost triple to invader in economical loss.

I think that premise needs nuance.  It seems to assume that an empire is comprised of equally valuable cities, all at an equal level of development, all equally well defended, sitting on equally valuable resources.

I recently watched a video about Hitler's decision to invade the USSR.  He made the claim that it was to get the oil.  That they were desperate for oil, that they were absolutely compelled to get oil, that their army was going to stop running if they didn't get the oil.  And that a lot of Hitler's generals didn't understand this, they had operational level ideas of going after Moscow or something... which would not have stopped the USSR from fighting.  Hitler didn't live through the war, his generals did, so they chose to blame the dead guy as "the madman".  Thereby hiding their own culpability for lack of understanding of economic objectives, and picking bad strategic targets.  So claimed this video.

We would like an AI that responds with rational economic interest.  That allocates its forces where they're going to hurt the enemy most, and knows how to cut its losses where it is weak.  This is probably not SMAC's AI.  :-)  So any game designs aimed at what we'd like to be true, need to be filtered through what is true about the AI behavior.

Or else one needs to write a new AI.  And for that level of work, I personally will write a commercial title I can make money from.

Now, if you're formulating designs based on human players vs. human players, all bets are off.  I wouldn't even have anything to say.  It's a damn long game and I've never actually slogged it out with a human who's my military equal.  I had a neighbor who was grossly inferior at everything, we played on a LAN a little bit back in the day.  I really couldn't tell you whether "triple economic damage" is a good way to design a human vs. human game or not.  It does sound arbitrary to me, potentially running the risk of an unsatisfying WW I type slog.

Back to the historical Hitler example.  His generals wanted to fight a war of maneuver, against the USSR.  The problem is, they didn't have the fuel to do the maneuvering...

> It seems that 50% is a sweet spot.

I like it just fine at that setting.  Failing to get rid of Sensors does hurt.  It is part of my conquest drill to deal with them.  At times they can seriously aid the enemy.  Only when I've reached the "I'm stomping you now" phase, does it stop mattering so much.  Even then it might still be advisable, for economy of blitz.

> Indeed, if you are 4 levels ahead you should be able to beat the crap out of anybody.

You have to actually manufacture this stuff.  You may not be 4 levels ahead in minerals or factories.  Advanced units are expensive units.

> I am unsure about needlejet cost.

I'm not sure either.  I went through a period them being helluh expensive.  This was not tenable.  Recently I've reduced costs.  I'm not sure if I need to reduce costs again.  I feel it needs playtesting and feedback from players other than myself.  I'm way too good at rail based invasions, to even care about air force.  The AI currently builds some planes, not many.  There are no clouds of planes like in the stock game.

> Same thinking about choppers.

Game mechanically, a unit that can fire every move, and has a high number of moves, does not belong in the game.  It is overpowered.  Many modders have said so, and have taken action to nerf it.  I tried reducing moves for awhile.  I eventually decided this made the unit unpleasant / unrewarding to consider building.  So then I removed it from the game entirely.  Game mechanically, I do not regret this.  Artistically, I regret the loss of the unit artwork.  I haven't come up for any answer to that, because I'm not interested with messing with art assets or other file hacks.  I seem to have a legal right to mod the *.txt files I actually did, and expanding that might nullify my rights.  For the amount of work I've put into my mod, I'm a bit conscious of those rights.  Feels like I should protect them, "just in case", unless there's something major to be gained by fiddling with more stuff.

[Alpha Centauri Bear]

> Conquering same technologically developed opponent bases should cost triple to invader in economical loss.

I think that premise needs nuance.  It seems to assume that an empire is comprised of equally valuable cities, all at an equal level of development, all equally well defended, sitting on equally valuable resources.

That's right. I meant "all other factors being equal" = about same economical power too. Besides, by the mid game faction technological development roughly corresponds its economical one. Specific details are irrelevant. I just want to avoid vanilla game cases when irregardless of economical advantage one could amass tons of needlejects and wipe out opponent in 5-10 turns without incurring any significant losses! Then, after conquering one faction, turn their indestructible army to next one producing more and more units meanwhile so indestructible army keeps growing like snowball.

> Indeed, if you are 4 levels ahead you should be able to beat the crap out of anybody.

You have to actually manufacture this stuff.  You may not be 4 levels ahead in minerals or factories.  Advanced units are expensive units.

That's true. What I meant is within the scope of my modding I am not going to restrict 4 levels more advanced faction from having indestructible army. They should deserve it at this point. As you said before: at some point your advantage should allow you to win the game already.

> Same thinking about choppers.

Game mechanically, a unit that can fire every move, and has a high number of moves, does not belong in the game.  It is overpowered.  Many modders have said so, and have taken action to nerf it.  I tried reducing moves for awhile.  I eventually decided this made the unit unpleasant / unrewarding to consider building.  So then I removed it from the game entirely.  Game mechanically, I do not regret this.  Artistically, I regret the loss of the unit artwork.  I haven't come up for any answer to that, because I'm not interested with messing with art assets or other file hacks.  I seem to have a legal right to mod the *.txt files I actually did, and expanding that might nullify my rights.  For the amount of work I've put into my mod, I'm a bit conscious of those rights.  Feels like I should protect them, "just in case", unless there's something major to be gained by fiddling with more stuff.

Multiple attacks sounds scary but in practice they are good against weakly armored units caught unprotected in a field. They are about equally pathetic against fully defended base with AAA units and AC. You'll lose somewhere 2-4 attackers per one defender. Your multiple attack ability won't do any good there. Sure if you already have weapon advantage that allows you to kill few defenders and survive, then copter just does job faster than needlejet that's all. Add to that its vulnerability to ground counterattack. If you don't get it back to base every turn it is most likely 100% dead. For me they are about as strong as planes just give you slightly different tactical usage.

[bvanevery]

The Needlejet swarm problem has a specific solution: make the Needlejet chassis more expensive.  The AI certainly toned down after I did that, and maybe I've gone too far.

I've seen way too many Chaos weapons mounted on Choppers to accept that they aren't doomsday weapons.