SMAX - The Will to Power - mod

[Alpha Centauri Bear]

It's interesting, that we see an opposite problem with AI

In my opinion without reactor advantage, it's not effective in the conquest enough and you think it's too effective already.

AI vs. human is completely different story.
If we are talking about AI vs AI both sides are equally skilled in combat, obviously. The problem is that both of them being quite sloppy it is often happens that one side that somehow amassed numbers already just beat the crap out of other who cannot organize proper defense. That being random you often end up with only one or two enormous factions those eat others by the mid game already. I am adding to defense to slow down such random beating to extent that most of faction can survive to the game end even if in shrink size.

That's all. I am slowing down conquest dynamic to match the game length.

[dino]

The problem is that both of them being quite sloppy it is often happens that one side that somehow amassed numbers already just beat the crap out of other who cannot organize proper defense.

It should be dealt with AI improvements, I have these ideas that I was thinking about trying to implement myself, but you can try too. It should took you 10 times less time then me.

The actual issues are:
1) thinker AI can sometimes remain completely without army even into midgame, if it has peace with neighbours and it's threat level was never ticked, no amount of unbalanced defense bonuses will fix that.
It's equally easy for a player, to steamroll such defended mostly by formers faction btw.
2) AI won't upgrade units,
3) Support level that gives free units up to the base size, is completely overpowered in the early and early midgame,

Solutions:
1) Modify thinker production code so it always produces some units up to certain producion percentage dedicated to support, regardless of the  threat level. I was thinking of making it configurable via thinker.ini for easy playtesting and tuning, I was thinking something like 35%( ini variable ) +- 5%* ai_fight ( so 30 - 40% depending on faction ). Tradeoff is 30% slower development, but AI would always be prepared for defense.
2) If 1) is implemented and player try to gain advantage by developing without any army ( more then 3x smaller army ), make *neighboring* ai_fight 0 and 1 factions without treaty change diplomacy status to vendetta ( with exception of early game ), or heavily deteriorate their disposition toward the player, if you could access this value.
3) Implement unit upgrade with money, in simillar fasion thinker implemented facilities rushing with money.
4) Modify SE settings and faction bonuses, so SUPPORT3 is not possible until MilAlg

By reading through thinker code, I believe that all of these can be accomplished with modifications to existing thinker codebase, without any additional reverse engineering.

It's not the combat odds that are real  reason for a problem you've identified, and such significant changes favoring defender odds won't really solve them, but will introduce other issues of their own.

[Alpha Centauri Bear]

Oh no, I didn't mean to solely protect human player with this. It is just game becomes pretty shallow when 4 out of 7 factions got eliminated early. I understand that someone has to win eventually just not that early.

Again. I am not defending my choice. Not much playtesting yet. I may be wrong.

[Alpha Centauri Bear]

1) thinker AI can sometimes remain completely without army even into midgame, if it has peace with neighbours and it's threat level was never ticked, no amount of unbalanced defense bonuses will fix that.
It's equally easy for a player, to steamroll such defended mostly by formers faction btw.

You are absolutely right but this is completely different issue and completely different fix not fixable by attack-defense balance.

2) AI won't upgrade units,

Wha? I never knew that.

3) Support level that gives free units up to the base size, is completely overpowered in the early and early midgame,

Hmm. I tend to agree but how this is relevant to attack-defense balance per se?

Solutions:
1) Modify thinker production code so it always produces some units up to certain producion percentage dedicated to support, regardless of the  threat level. I was thinking of making it configurable via thinker.ini for easy playtesting and tuning, I was thinking something like 35%( ini variable ) +- 5%* ai_fight ( so 30 - 40% depending on faction ). Tradeoff is 30% slower development, but AI would always be prepared for defense.

Makes sense.

2) If 1) is implemented and player try to gain advantage by developing without any army ( more then 3x smaller army ), make *neighboring* ai_fight 0 and 1 factions without treaty change diplomacy status to vendetta ( with exception of early game ), or heavily deteriorate their disposition toward the player, if you could access this value.

I think something like that is already present in the game. The weaker you are the bolder the AI. It doesn't happen instantly. They go through threats and extortion first then to breaking pacts, etc. To get strong piece human has to not attack first and have strong army. Besides, factions become generally more and more aggressive with time. They attack you sooner or later anyway. Besides or the above two it is insanely difficult to implement (properly). I am sure I won't bother with it.

3) Implement unit upgrade with money, in simillar fasion thinker implemented facilities rushing with money.

AI didn't rush facilities either? What do they do with all these money?

I guess this is a good TODO.

4) Modify SE settings and faction bonuses, so SUPPORT3 is not possible until MilAlg

I think you overestimate its value. Yes it's good but not that insanely. For small early game bases "4 or up to base size" essentially means just 4. For medium size best unit producer game bases have 10-20 mineral production. The difference between 4 and base size is probably somewhere 2-4 at best and supporting 2-4 extra units is not that hurtful comparing to base mineral production.

It's not the combat odds that are real  reason for a problem you've identified, and such significant changes favoring defender odds won't really solve them, but will introduce other issues of their own.

Yep. Combat odds is just one of the problem. I was disabling human indestructible snowball army exploit by this. Unfortunately, there is no other way to do it without making base defense noticeably stronger to turn cakewalk into a hard battle. So I suggest this change (attack-defense balance) stays at least to some extent and we'll work on other issues meanwhile.

[Alpha Centauri Bear]

Found it.
Here is the deal, dino.
PD and TF multipliers can be directly changed by 50% increment only. Let me know if you like to try this.
Another option is to completely override the combat calculation function. This way I can change outcome in a more granular way. Somewhere in 1/0x100 or something. However, overriding whole function means there are a lot of conditions and variables I haven't even see = a lot of different use cases and testing to make sure I didn't screw them up by changing just one case. Are you up to it?

[Alpha Centauri Bear]

Implemented first option in version 22. Enjoy!


I don't think there is a need for less than 50% step granularity but you'll tell me after tests.
Intrinsic base defense is in alphax.txt if you need to tweak it as well.

[dino]

Thank you, I'll play with your mod through the weekend.

I'll try to first give a shot to the values you've set and if I indeed encounter the issues I foresee, I'll start doing other experiments. With 50% granularity I'd go for PD 100% and TF 50% probably if had to change it.
As for skipping collateral from defeated  unit, there is a check to skip it if inside the base, it should be enough to modify its condition and artillery percentages are already avaiable in vanilla alphax.txt which I forgot.

I was disabling human indestructible snowball army exploit by this.

I think if Inductio implemented magnets only usable on owned territory and gave AI a magnet network, it would be the best counter to the player snowballing.
Since it wouldn't be possible to conquer in a single turn more then the outer ring of less developed bases, which could be met with a retaliation of the whole remaining AI army in the next turn.
If you just increase defense rating to the point even player will struggle with conquest, how can you expect AIs to deal with it during their offensives ?

[Alpha Centauri Bear]

I'll try to first give a shot to the values you've set and if I indeed encounter the issues I foresee, I'll start doing other experiments.

I used vanilla values for default in thinker.ini, of course.

If you have plenty of time to test few games up to TF then sure do few games for comparison. Otherwise, I'd encourage you to test your theory first to get a taste whether it is something you expected or it is too much.

[Alpha Centauri Bear]

As for skipping collateral from defeated  unit, there is a check to skip it if inside the base, it should be enough to modify its condition

Thanks for pointer. Will look into it.

I still cannot wrap my head around why you want to play with it? This one seems to be nicely designed and causes least problem. It is a natural defense against stack of doom when attacker can move ~100 unit stack of mixed attackers and defenders to the base and be sure none of their attacker will be even slightly damaged by preventive strikes from the base. Decreasing collateral damage from 100% in Civ 1/2 to 20% seems like a brilliant idea. Now you still can destroy whole stack but you need 5 successful attacks against it. If less then stacked attackers can still attack in their turn but with lower power.

[Alpha Centauri Bear]

Tested version 23 with new unit hurry and upgrade cost formulas.
Negative cost for reactor upgrade works! Cool.
Overall both costs seem a bit not affordable for frequent use especially in early game. This is fine for hurry cost as they were not affordable before too. Upgrade is higher for lower tier units and lower for higher tier so OK too.

[Alpha Centauri Bear]

Found an exploit. Refund for upgrade to cheaper unit is 4x minerals difference. That allows building more expensive units with obsolete reactor, upgrade them and use the refund to buy buildings twice as cheap.
I should limit refund with x2 or x1.

[Alpha Centauri Bear]

v23 with collateral damage parameters.

[dino]

I still cannot wrap my head around why you want to play with it? This one seems to be nicely designed and causes least problem. It is a natural defense against stack of doom when attacker can move ~100 unit stack of mixed attackers and defenders to the base and be sure none of their attacker will be even slightly damaged by preventive strikes from the base. Decreasing collateral damage from 100% in Civ 1/2 to 20% seems like a brilliant idea. Now you still can destroy whole stack but you need 5 successful attacks against it. If less then stacked attackers can still attack in their turn but with lower power.

For this reason I use only large stacks of speeders, or elite infantry, so I can move my stack and attack in the same turn with full strenght and without a risk of loosing from retaliation if I successfullly conquer the base. AI doesn't do this and their stacks are free game, so it's another nerf that should benefit AI much more than the human player. It doesn't also makes sense, why should in lets say a 25 unit stack, killing 20% units completely doom remaining 80% ? It's not even necessary, with the way multiround combat works, even killing just 2 units in a stack renders remaing force completely useless for a base assault. It's a feature for a casual player imo, that prevents him from getting bored, evolution of which bacame 1UPT which ruined the CIV series. Of course such a huge change may break the game in a way that I can't foresee yet, but I always wanted to try it.

I hope you didn't loose too much time doing me these favors. I don't have any more of such ideas, so hopefully I won't bother you with more tasks

[Alpha Centauri Bear]

I still cannot wrap my head around why you want to play with it? This one seems to be nicely designed and causes least problem. It is a natural defense against stack of doom when attacker can move ~100 unit stack of mixed attackers and defenders to the base and be sure none of their attacker will be even slightly damaged by preventive strikes from the base. Decreasing collateral damage from 100% in Civ 1/2 to 20% seems like a brilliant idea. Now you still can destroy whole stack but you need 5 successful attacks against it. If less then stacked attackers can still attack in their turn but with lower power.

For this reason I use only large stacks of speeders, or elite infantry, so I can move my stack and attack in the same turn with full strenght and without a risk of loosing from retaliation if I successfullly conquer the base. AI doesn't do this and their stacks are free game, so it's another nerf that should benefit AI much more than the human player.

Exactly, man! For this reason. That is the whole point of any rule in the game - to counter some other tactics.

It doesn't also makes sense, why should in lets say a 25 unit stack, killing 20% units completely doom remaining 80%

You are right. Real life works differently. But this is irrelevant to game design. Let's avoid such arguments.

? It's not even necessary, with the way multiround combat works, even killing just 2 units in a stack renders remaing force completely useless for a base assault. It's a feature for a casual player imo, that prevents him from getting bored, evolution of which bacame 1UPT which ruined the CIV series. Of course such a huge change may break the game in a way that I can't foresee yet, but I always wanted to try it.

Sorry, I don't follow what ruined Civ series? The ability to kill all units in square with one attack? I thought this was a fresh innovation to introduce new distinctive tactical element in game comparing to predecessors. I recall it awed me in Civ1.

I hope you didn't loose too much time doing me these favors. I don't have any more of such ideas, so hopefully I won't bother you with more tasks

Not at all. I like follow ideas and see what happens. This is all free time leisure anyway.



At the same time it is better to have some sort of focus in mod. Otherwise, fixing never ends. That is why I am asking these stupid questions about whether some change is needed. Please forgive me.
For example, I can understand that stronger armor + TF  may lead to conquest stalemate. That hurts aggressive factions and tilts balance to economical ones. This need to be tested, of course, but I understand there could be potential problem.
With collateral damage I don't see a problem. Not that it doesn't exist - I just cannot understand if it exists. So I'm happy to give you all the configuration parameters to try and prove your point.

[Alpha Centauri Bear]

Here is another example of my reasoning. Recently I've removed all hurry restrictions and made all hurry cost flat.

Vanilla implementation: different formula for facilities and units, different multiplier for regular facilities and SP, formulas use penalty thresholds, thresholds are different for regular facilities, unit, and SP. That results in player read formulas in help or internet, count tiny squares on the screen with pencil to get number of rows and accumulated minerals, use pen/paper/calculator to come up with answer, use knowledge of carry over minerals to not overflow current production, and on and on. This all distracts from actual playing the game.
The only meaningful rule here is to price SP higher than regular facilities. That is understandable because SP are one of a kind and there is always race to build it. Therefore, higher price limits money bags from buying it outright using previously accumulated reserves. In other words, it assures SP building does not happen instantaneously so other factions can participate in race.
Other rules are just garbage. Obviously, no one has that much money to buy more than 90% of all production. One may buy Recycling Tanks in one base but not in another. Thresholds forces player to not buy stuff when it doesn't have 10 minerals invested in yet. What does it change? Now player will just pay for half of Recycling Tanks in both bases. Same amount of money spent = same amount of mineral production substituted. No economical impact. Sure that prevents player to develop a single base very quickly using whole empire resources. What is the application of this? Building drone facilities in captured bases? In this case it would be more logical to penalize rushing in captured bases specifically and not to introduce rule affecting the rest of empire.
Same story with unit rushing. Beyond 6 rows of minerals the cost becomes exorbitant and keep growing even if no one ever spend money on it anymore. I don't see a point making later game units much more expensive. At any game stage the unit you rush is the most important unit helping you to organize defense or something. Just make them all expensive and be done with it.