SMAX - The Will to Power - mod

[Nexii]

I mean you only need like 2 units a base unless you do some rebalancing of combat and/or unit costs. Raise up ECM and AAA power and you'll have to artillery and/or probe a base down before taking it. Combined arms should be needed but in the stock game speeders (and later needlejets, copters) are so fast and cheap that nothing else stands a chance.

Similarly sea units play more of a role when they don't automatically die to air

[bvanevery]

What support really influences past the early game in vanilla is how many stock formers a base can employ.

I have stock AI making use of Clean Formers since the beginning of the game.  Anecdotes from playtesters is that my AI factions terraform better than stock.  Unsurprising as the whole intent was to free up the AI's SUPPORT burden.  Stock AI just doesn't know what to do about SUPPORT, and would otherwise exhaust itself.

[Alpha Centauri Bear]

Fuf. Finally released AI update. Just tired.
😪

It was mostly inspired by latest playtest revealing AI as not competitive. I cannot say it is there yet but the big number of modification are done. Check changelog for short list of changes.
Usual warning: major code rework. I think I squashed all the bugs but nobody knows. Don't use for multiplayer campaigns yet.

Here is the more elaborated discussion.

I was focusing more on development side: expansion, growth, production. Just because it was there already and much more easily quantifiable and measurable.
* Big improvement in base production and placement. Now every tile is evaluated. AI searches for more lucrative locations.
* Big/Huge/Enormous improvement on land unit transportation by sea. Generalized algorithms allows seamless transitions all over the globe. AI is no more technically challenged to jump onto/into another continent/ocean. All factions exploit land and ocean alike.
* Added more sophisticated algorithms to production choices. For example, TECHSHARE faction values labs facilities less and other minor staff here and there.

There is one improvement on army production and movement, though. AI evaluates base threats. That is whether base can sustain nearby enemy military attack should they move onto. Current diplomatic relations are taken into account. Current enemy represents full threat. Whereas peaceful factions just partial to be prepared for sudden attack. AI uses this information to build and distribute protectors across bases. It also tries to match them with most effective counter units. It also tries to produce best counters available. The computations are insanely heavy as they need to account for every base and every unit and also cross unit-unit effectiveness. So, at this point, it is very very crude. I have tested it for the first 100 turns so far but not beyond it. Need to verify AI is doing sane things with bombers, interceptors, natives, anti-natives, ships, etc. And especially artillery as it the most difficult part to evaluate.

At the same time it does build sensors, Perimeters in frontier bases, Command Centers, and the like. Also overall military production priority is raised very high. It should be able to consolidate and fend off aggressors if needed. But again - needs a lot of testing as combat is the most complex activity in this game and I don't even pretend to approach perfection.


In conclusion here are some power graphs attached. I started two games from the same initial placement: University (single base human observer), Pirates, Morganites, Angels, Gaians, Drones, Cyborgs. Then I enabled factions 2,3,4 to use WPT algorithms leaving factions 5,6,7 to Vanilla/Thinker (left chart). In second game I swapped them (right chart). 100 turns. That is not to compare WTP with Thinker but to establish some average level and verify whether WTP early expansion/development indeed improved in this version.
My general feeling is that WTP lead AI is not always better but it never falls too much below average. Where previous version AI could get stuck in pretty bad position, new one finds ways to escape to ocean, other continents, proper terraforming, etc.

[bvanevery]

So you are in essence testing 6 player games.  That is not necessary.  In Scenario Editor: start playing a faction manually as a human, i.e. the University.  Play the 1st turn when you land.  End your turn.  2nd turn, switch your view to a different faction.  Choose to passively observe, don't take over the faction.  Now you've got 7 AIs playing each other.  The AI you started with, will not be "fully automated" on that 1st turn, but it's an acceptable price to pay.

Apologies if I don't actually have time to look at your work for awhile.  I recently bought a brand new laptop with a RTX 3060 in it.  I've made no use of that 3D graphics capability at all, so far.  Verified I'm getting the expected performance with a few benchmarks, but that's it.  No games.  I'm only barely starting to get used to the new keyboard and touchpad gestures.  Haven't compiled anything yet.  Life is just chaos right now.

[Alpha Centauri Bear]

Apologies if I don't actually have time to look at your work for awhile.  I recently bought a brand new laptop with a RTX 3060 in it.  I've made no use of that 3D graphics capability at all, so far.  Verified I'm getting the expected performance with a few benchmarks, but that's it.  No games.  I'm only barely starting to get used to the new keyboard and touchpad gestures.  Haven't compiled anything yet.  Life is just chaos right now.

Please don't apologize. It's not that you (or anybody else) own something. Not even me to myself.
😂

You helped a lot already. In big part this activity is due to you previous tests.

[bvanevery]

Long live the Issue Tracker!   

 

[Sardaukar]

Will to Power is great. I'm still using an older version (167?), with some mods to the txt. Noticed there's an optional feature where you can set roads not to work in enemy territory. Didn't use it because seems kinda hokey. Why would roads stop working because they're on the other side of invisible line on a map? This would also really change up the meta.

Had a new thought about this today. The road improvement represents not just pavement but logistical support for transit. Naturally, this will all shut down in the face of hostile forces moving up the highway. So there's the justification.

However, road as movement multiplier seems a sketchy game mechanic. Yes, some units are infantry but these are probably small military kill teams, not WWII divisions on the eastern front. We can reasonably imagine high tech warfare as tiny forces applying tremendous force against key points in the enemy's infrastructure. Infantry units would fight on foot but they'll travel (especially on their own roads) in vehicles like everyone else. Who in his right mind would walk across hundreds of km of unbreathable Chiron landscape from Baikonur Cosmodrome to University Base?

With that in mind, roads providing a multiplier is funky. Rovers and infantry (riding rovers that belong to basic faction infrastructure), should move at about the same speed. Instead of a multiplier, why not a standard movement bonus to roads?

For example, all ground units may move over 10 friendly road squares for free. Could be some value other than 10 but I like the thought of rapid redeployment around your territory. Plus, this would ensure that hovertanks and rovers don't move less than in vanilla. It would, of course, substantially alter the meta but that could be resolved through testing and fine tuning a number of variables.

However, it's also possible to imagine bringing up your own logistics to make use of enemy roads. Thinking of a special unit with a terraform (?) ability to convert enemy territory into your own. Logistical Ops Squad. Could use the currently ignored supply crawler weapon.

[Alpha Centauri Bear]

Thank you for feedback!

Road usage switch is probably disabled due to the fact that multiple reviewers could not find a common ground whether to accept it or not. Too many disputable points as you mentioned yourself.
It seems that it is not too much problem for attacker to contain the invasion on land anyway.

[bvanevery]

I ended up uninstalling Galactic Civilizations III after about 80 hours of play, because I saw no evidence of the AI being able to fight in the early to early midgame.  Higher levels of difficulty would only add cheating movement and money buffs to the AIs.  That was boring.  So, now I'm trying WTP again.

Version 292.  terranx_mod.exe seems to ignore DirectDraw=0 in my Alpha Centauri.ini file.  terranx.exe honors it just fine.  I've got a higher resolution laptop screen to look at now, 1920x1080.  Also it's Windows 11.  I don't know what previous versions of the mod did.  I was using a different laptop then.

I put that setting in my Alpha Centauri.ini but did not change thinker.ini.  It has the following settings:

Code: [Select]

; Make SMACX run on the native desktop resolution by disabling DirectDraw.
; Recommended value is 0. This overrides any value in Alpha Centauri.ini.
DirectDraw=0

; Enable windowed mode in custom resolution while DirectDraw=0 is set.
; If windowed mode is disabled, native desktop resolution is used instead (default).
; This mode can also be enabled from the command line with parameter "-windowed".
windowed=0
window_width=1024
window_height=768

I don't think it's using native desktop resolution.

Changing window_width=1920 and window_height=1080 doesn't do anything.  If I also change windowed=1, I get this totally weirdo blown up window where I can only see 1/4 of the menu screen.  Can't access the menu buttons so I just hit ESC to quit.

Choosing windowed=1 and the defaults of window_width=1024 and window_height=768, results in a borderless window stuck in the upper left corner of my desktop that cannot be moved.  Running off the border of the window to the left or top will scroll the window, but going to the right or bottom will access the desktop and pass out of the game.  This feels quite broken.

One might think windowed=1 and window_width=1080 and window_height=1080 would do something logical, like make a square on the left side of the screen.  However the Windows 11 taskbar overrides the bottom of the window, and the menu background art is also stretched.  I can't access the menu buttons, so this pretty much doesn't work.

windowed=1 and window_width=1024 and window_height=1024 is about the same.  Basically windowed doesn't accomplish anything.

Ok, the problem has something to do with Windows 11 scaling behavior.  I found a solution.  Right click Will To Power's shortcut on the desktop (that's how I installed it for myself) and click Properties.  Select Compatibility.  Select "Change high DPI settings".  A section at the bottom is titled "High DPI scaling override".  It has a checkbox, "Override high DPI scaling behavior".  Check it, and in the drop down menu, select "Scaling performed by:" Application.  Now you will get SMAC in teeny weeny 1920x1080 glory.

[Sardaukar]

Thank you for feedback!

Road usage switch is probably disabled due to the fact that multiple reviewers could not find a common ground whether to accept it or not. Too many disputable points as you mentioned yourself.
It seems that it is not too much problem for attacker to contain the invasion on land anyway.

Yeah, the counterarguments are strong. But what about this?

With that in mind, roads providing a multiplier is funky. Rovers and infantry (riding rovers that belong to basic faction infrastructure), should move at about the same speed. Instead of a multiplier, why not a standard movement bonus to roads?

For example, all ground units may move over 10 friendly road squares for free. Could be some value other than 10 but I like the thought of rapid redeployment around your territory. Plus, this would ensure that hovertanks and rovers don't move less than in vanilla. It would, of course, substantially alter the meta but that could be resolved through testing and fine tuning a number of variables.

Would be more realistic, make for faster gameplay and alleviate that 4x weirdness where wars take way more years than they should. To match the increased road speed, would also want to increase movement for ships and possibly air units but that's easy to do with the txt. Increased movement would make offense far stronger but the territorial defense bonus you added could cancel that out. Either that or make the base/perimeter defense stronger.

[Alpha Centauri Bear]

I ended up uninstalling Galactic Civilizations III after about 80 hours of play, because I saw no evidence of the AI being able to fight in the early to early midgame.  Higher levels of difficulty would only add cheating movement and money buffs to the AIs.

Sadly, this can be said about ANY game pre 2000. They didn't do good AI then.

Thank you for resolving this issue with resolution on Windows 11. I don't have it yet but it's good to have it solved already!

[Alpha Centauri Bear]

Yeah, the counterarguments are strong. But what about this?

With that in mind, roads providing a multiplier is funky. Rovers and infantry (riding rovers that belong to basic faction infrastructure), should move at about the same speed. Instead of a multiplier, why not a standard movement bonus to roads?

For example, all ground units may move over 10 friendly road squares for free. Could be some value other than 10 but I like the thought of rapid redeployment around your territory. Plus, this would ensure that hovertanks and rovers don't move less than in vanilla. It would, of course, substantially alter the meta but that could be resolved through testing and fine tuning a number of variables.

Would be more realistic, make for faster gameplay and alleviate that 4x weirdness where wars take way more years than they should. To match the increased road speed, would also want to increase movement for ships and possibly air units but that's easy to do with the txt. Increased movement would make offense far stronger but the territorial defense bonus you added could cancel that out. Either that or make the base/perimeter defense stronger.

I agree that would be more realistic to imagine that foot soldiers are actually transferred by some land transport. This would require complete game rewrite, though. Currently units have only their speed and road multiplier (which is common parameter). To implement your proposal we would need to keep "speed without road" and "speed on road" parameters and there is no space for them both, unfortunately. It is too much game mechanics change proposal. Would you be willing to invent something of the same sort but within the current game mechanics scope? Probably by changing some parameters only?

Besides, what would be the difference between land classes chassis then if infantry start moving that fast?

[Sardaukar]

You know the code and I don't but I imagined something along the lines of a unit having 10 "ticks" for free movement between two road tiles. When all ticks were exhausted, it would have to use ordinary movement. The game already has "ticks" or whatever they're called to keep track of how much movement a unit has already used in the turn. Magtubes factor into this somewhere, as they simply don't deplete "ticks."

But I think that's what your describing as not possible/inordinately difficult.

Rovers would still be slightly faster on roads (10+2 rather than 10+1) but mostly they'd be advantageous for off-road operations. Suppose it might be appropriate to increase their speed as well.

If it could be made to work, would eliminate 2/3 and 1/3 strength penalties.

[Alpha Centauri Bear]

You know the code and I don't but I imagined something along the lines of a unit having 10 "ticks" for free movement between two road tiles. When all ticks were exhausted, it would have to use ordinary movement. The game already has "ticks" or whatever they're called to keep track of how much movement a unit has already used in the turn. Magtubes factor into this somewhere, as they simply don't deplete "ticks."

But I think that's what your describing as not possible/inordinately difficult.

One doesn't need to know the code to see values/parameters in the game. Like infantry chassis can move 1 open terrain tile while rover can move 2. These values can easily be changed even in alphax.txt. I meant these kind of changes. Something that can be done through text modding mostly.

The thing is that each unit has a single variable that counts its movement allowance. So introducing another movement pattern would require second counter and some interchanging between two. And there is no place in vehicle structure to fit second one into.

Here is some info for you to play around.
There is a chassis speed which you can see in alphax.txt. This is how fast they move on open ground without road. Value can be simply edited in file.
There is a "road/tube multiplier". It is 6 in current WTP version but can be modified in alphax.txt and thinker.ini.
Chassis speed times tube multiplier = total number of "movement points" which unit can spend. The value should be whole.
The cost of movement to next hex is defined by the terrain (open, rough, fungus, road, tube). This cost should be also whole!!!
Normally moving to open terrain without road/tube/river costs exactly the multiplier so number of hexes unit can move in open properly divides down to chassis speed.
Last step can be done even if there are insufficient movement cost.

Example.
Rover. Chassis speed = 2. WTP road/tube multiplier = 6. Total movement points = 2 * 6 = 12.
Cost to move to open terrain (no road/tube/river) = 6. Therefore rover can make 12 / 6 = 2 steps in open.
Cost to move on road = 2. So it can move 12 / 2 = 6 hexes along the road.
Cost to move on tube = 1. Same: 12 / 1 = 12 hexes along the tube.

Would it be close to your idea if we just increase movement allowance to all land units by fixed value? Currently they have 1-2-3 (infantry, rover, tank). What if we set them to: 2-3-4 or 3-4-5, for example? That will proportionally increase their road speed, of course.

Another easy option I see is to change movement cost along the road based on chassis type and make infantry to move on roads as if they are tubes. Or we can do other way around: make rover moving x1.5 or x2.0 times slower on roads implying that they are already mobile units and use their own motor power moving on roads. Whereas infantry is using transportation that increase their road speed tremendously. Hovertank would benefit road the least as it already don't have impassable terrain so the road doesn't add much to it.

Something like that?

chassis	chassis speed	movement points	field cost -> distance	road cost -> distance	tube cost -> distance
infantry	1	6	6 -> 1	2 -> 3	1 -> 6
speeder	2	12	6 -> 2	3 -> 4	2 -> 6
hovertank	3	18	6 -> 3	4 -> 5	3 -> 6

Although, many people would be quite disappointed by road movement speed reduction for fast units.

If it could be made to work, would eliminate 2/3 and 1/3 strength penalties.

That's doable. However, it gives enormous advantage to infantry. They are now as good as speeder attacking from afar.

[bvanevery]

Sadly, this can be said about ANY game pre 2000. They didn't do good AI then.

GalCiv3 actually came out in 2015.  The last expansion, Retribution, was in 2019.  That's what I just played.  Now they just shipped GC4.  I don't rule out the possibility of trying 4, but with the unimpressive AI performance of 3, it's not going to be 1st in line.

GOG has got Old World now, with a 10% sale on it.  I might do it.  But... today I thought, let's do WTP first.  And I'm somewhat interested in the AI of Remnants of the Precursors, the one Xilmi made.

Thank you for resolving this issue with resolution on Windows 11. I don't have it yet but it's good to have it solved already!

Definitely some wisdom to spread around.  Might need an entry in the game's installation notes, if it proves to be a problem for more people.