Unit cost change ideas

[Yitzi]

A discussion in another thread has brought up the fact that the unit cost formula is a major contributor to certain imbalances, and needs fixing.  Of course, this would all need to be made optional, so here's where we'll provide our ideas for changes we think would be good.  Each change should mark which of the previous changes it is compatible with (i.e. you can do both) and which it requires in order to work as intended; if two or more changes are meant to work together, list them as a single change.  All fractions are assumed to round down, except for those involved with the reactor (where they round up if more than 50% and down if less than 50%).  So I'll start; all of these are compatible with each other:

1. Infantry cost is halved-and-then-add-one after applying the minimum reactor cost, rather than before.
2. The minimum weapon cost is changed from armor/2+1 to (armor+2)/3.
3. Reactor effects on cost are changed; instead of dividing by 2^(1+R), multiply by (6-R) and divide by 20. Thus, instead of dividing by 4,8,16,32, it divides by 4,5,6.66...,10.  (The increasing proportional discount is intentional, to compensate for the decreasing proportional hitpoint increase.  Thus, the total effectiveness increase will be 2.5 with Fusion over Fission, 2 with Quantum over Fusion, and 2 with Singularity over Quantum, instead of 4, 3, and 2.66.)
4. Minimum reactor cost is decreased to (R+1) rows.  (Basically, quantum and singularity reactors get a bit cheaper).  This requires 3, and will work best with 1 and 2 as well.
5. Noncombat units do not have their cost increased for having both module and armor cost greater than 1.

[Nexii]

Reposting this, I'd like more people's ideas/critique.

A much simpler and intuitive unit cost formula would take reactors out of the equation, and make it so better weapons/armor always provided better efficiency.

Proposed cost formula:

M = (W + A + C) * 10

M: the cost of the unit in minerals
W: weapon factor.  would scale from 0 for early game weapons to 6 for the late-game weapons.
A: armor factor.  would scale from 0 for early game armors to 4 for the late-game armors.
C: chassis factor.  1 for infantry, 2 for rover, 3 for hovertank, 4 for needlejets and gravships, 5 for copters. 1 for foil and 2 for cruiser.  (these are debatable)

Since weapons have double the values of armor, and a 3:2 cost ratio, this would give a 3:2 effectiveness ratio of attack to defense (the same as PSI).

For rovers, hovertanks, foils, cruisers, needlejets, copters, and gravships, add (W*A) to the above cost.

AAA would likely be put to Cost: 0 as well under this model.

[Nexii]

I like #5 as a change.  Commenting more on #3 here.  A 2.5x increase in cost effectiveness at Fusion is still a lot.  I see that in my formula, the cost effectiveness increases would be 2, 1.5, and 1.3333 This isn't a constant...it's a good point that each reactor upgrade should be.  A thought I had was to instead give each reactor a proportional HP increase.  This would eliminate the need for a dividing cost factor that changes with reactor.

Considering that reactors increase both weapon and armor, what if we could make it so each new Reactor was only ~1.5x increase in cost effectiveness?  Even 50% is a lot and more balanced, I feel.  A doubling to military cost effectiveness is quite game-breaking.  As you pointed out, Fusion is currently a 4x cost effectiveness increase which really illustrates how overpowered reactors are.  I also feel that better weapons and armor should make troops more cost effective...it shouldn't be entirely on the (nerfed-down) reactors.  Otherwise the only benefit is that better weapons and armor give your troops a better chance to live in a single battle.  While this is nice, it typically isn't really a compelling enough reason on its own to pursue the upgrades.

[Yitzi]

Commenting more on #3 here.  A 2.5x increase in cost effectiveness at Fusion is still a lot.

Yeah, well...most of that comes from doubling hit points.  The fact is, reactors are powerful, and that's probably ok, because if you're getting reactors then the other guy might be using the time to grab choppers or magtubes or genejack factories.

A thought I had was to instead give each reactor a proportional HP increase.

So maybe the first would have 8 hit points, the second would have 12, the third would have 18, the fourth would have 27?  That'd be a pain to code, though.

I also feel that better weapons and armor should make troops more cost effective

They are, due to healing (or repair if you prefer).  A 6/1/1 attacker can kill dozens of 1/3/1 defenders, if he spends a turn in a base with a command center (or even a few turns in the back lines) between each one.

[Nexii]

The ability to repair is important I will agree on that.  As well your troops gain ranks when they do not die, that's also an important consideration.

I suppose I should clarify.  What I mean is that upgrading the weapon/armor should never result in a negative to cost efficiency.  Being on par might be acceptable.  In my opinion, there should be a slight cost efficiency increase for better weapons and armor (though much less so than reactors).  For example, an Impact Rover is 30 minerals (7.5 minerals/attack).  A Missile Speeder is 50 (8.333 minerals/attack).  I think this is the same issue I had in my formula, when armor and weapon cost factors start at 1 rather than 0 (as they should).

Reactors could double I suppose (10,20,40,80) with no cost reduction.  That would leave them strong.  I think that if you're going to go to all the work to re-code unit costs, I would do it with as general a formula as possible.  That way, it's easier to tweak the corresponding values to your taste in alphax.txt.  Let people play with various air cost factors and such.  The issue I see with keeping reactor cost reductions (rather than scaling up HP only) is that air quartering cost reduction.  Air chassis become very cheap at later reactors because of this.  What if instead, the chassis was a cost multiplier applied after all the combined weapon/armor cost was calculated?  It could work more like adding modifiers to a unit.  I think its needless complexity to have different rules for air (quartering), sea (halved), infantry (halved after min cost applied, then +1)   The problem I forsee with my formula is that with a chassis as a constant, better chassis becomes cheaper relative to infantry at higher weapon/armors as well.

It's a side thing, but I'd also prefer combat was speeded up for later reactors.  That might also be too much to change?

[Yitzi]

I suppose I should clarify.  What I mean is that upgrading the weapon/armor should never result in a negative to cost efficiency.

Firstly, that's extremely difficult to pull off exactly, due to rounding issues.  Secondly, in the particular case of the String Disruptor, the even-higher-than normal cost is justified because the absurdly high cost is its balancing factor (as it's far more powerful than defenses get).

For example, an Impact Rover is 30 minerals (7.5 minerals/attack).  A Missile Speeder is 50 (8.333 minerals/attack).

This is really a rounding issue.  Really, the missile speeder should be 45, but it can't be partial-rows, so it rounds to 50.  Like most rounding issues, it's not going to make a particularly large difference; here it's an 11% decrease, which should be less significant than the ability to attack a plasma garrison with a perimeter defense and have a good chance of winning.

Reactors could double I suppose (10,20,40,80) with no cost reduction.  That would leave them strong.

But it would also leave units absurdly expensive.

I think that if you're going to go to all the work to re-code unit costs, I would do it with as general a formula as possible.

The problem is, I don't see any easy way to generalize all the ideas here.  They're essentially fundamentally different ideas, rather than tweakable changes.

Let people play with various air cost factors and such.

It's an idea, but I'm not sure that it's worth the trouble when it can already be approximated by chassis cost changes.

The issue I see with keeping reactor cost reductions (rather than scaling up HP only) is that air quartering cost reduction.  Air chassis become very cheap at later reactors because of this.

Not really; 8 chassis cost is quite significant, so they still stay around speeder cost.

The problem I forsee with my formula is that with a chassis as a constant, better chassis becomes cheaper relative to infantry at higher weapon/armors as well.

Indeed.

It's a side thing, but I'd also prefer combat was speeded up for later reactors.  That might also be too much to change?

Speeded up as in the animation?  I think you can turn the animation off if you want.  (It would be too much to change easily.)

[Nexii]

It's similar idea (either way, you get a more linear increase from reactors).  Perhaps it is easier to change reactor reduction than change HP.  I would put the reactor factors down a bit from 2x, if feasible.  Consider that reactors also have the same effect that a weapon/armor increase does in helping a unit survive another battle, plus it gets the cost reduction applied.  I don't know...I just never really get excited about a weapon/armor increase the way I do when I get a new chassis or reactor.  It feels like weapon/armor upgrades only matter really early when support costs mean something.

It seems to me that there were more extreme examples of strange cost jumps at later reactors than with Fission.  I'd have to run some examples and see if this is the case.  Logically it should be less where rounding factors get smaller. 

I don't think more costly and more powerful late game units would be so bad.  A lot of common units are 60 minerals currently (and less, with + IND, and the proposed change to infantry cost), whereas late-game cities often produce 100+ minerals a turn.  I think it's more interesting if costs and power rise more...otherwise it comes down to 'the best 1-turn unit you can make, without waste'.

Although, something would have to be done with Native life.  But Native life / PSI units have their own issues.

[Yitzi]

It's similar idea (either way, you get a more linear increase from reactors).

Yes, except that if you don't have cost reduction, it gets a bit too expensive.

Perhaps it is easier to change reactor reduction than change HP.  I would put the reactor factors down a bit from 2x, if feasible.  Consider that reactors also have the same effect that a weapon/armor increase does in helping a unit survive another battle, plus it gets the cost reduction applied.  I don't know...I just never really get excited about a weapon/armor increase the way I do when I get a new chassis or reactor.  It feels like weapon/armor upgrades only matter really early when support costs mean something.

Firstly, support costs mean something later on*, though perhaps not as much.

*Well, unless you're making everything clean, but that's a lot more expensive and probably isn't worth it in a serious fight unless you're producing only a small amount of minerals per base...and why would you be doing that in a serious fight?

It seems to me that there were more extreme examples of strange cost jumps at later reactors than with Fission.  I'd have to run some examples and see if this is the case.  Logically it should be less where rounding factors get smaller.

Let me know what you find.

I don't think more costly and more powerful late game units would be so bad.  A lot of common units are 60 minerals currently (and less, with + IND, and the proposed change to infantry cost), whereas late-game cities often produce 100+ minerals a turn.

It is a concern, but changing reactor worth is too difficult.  I'll have to think about it.

Also be aware that once you get into the thousands for unit cost, you do start running into rollover errors.  (Not sure if it'll happen as early as 1280, or not until 2560.)

I think it's more interesting if costs and power rise more...otherwise it comes down to 'the best 1-turn unit you can make, without waste'.

Probably, although putting both weapon and armor on something is a good way to get expensive units.  But even so, single-focus units should still at least use up one turn's production on a good base.

Although, something would have to be done with Native life.

I believe native life effectively ignores reactors anyway.

[Yitzi]

Here's an idea to deal with the "best one-turn unit" thing: Keep the cost formula, but increase the value (and thus cost) of weapons.  Let's divide weapons and armor into "tiers":
Tier 0 is gun/unarmored.
Tier 1 is laser/synthmetal.
Tier 2 is impact/gatling/plasma.  (Gatling gets lumped with impact despite being better and a bit higher on the tech tree because it's on the way to Fusion and that makes it more likely to be gotten.)
Tier 3 is Missile/{whatever advanced subatomic theory will give according to my plan}
Tier 4 is Chaos/Silksteel
Tier 5 is Fusion/Photon
Tier 6 is Tachyon/Shard/Photon-with-tachyon-fields-available
Tier 7 is Quantum/Neutronium/psi
Tier 8 is graviton/antimatter.
Tier 9 is singularity/stasis
Tier 10 is String.

Now we can set a goal of the approximate increase in weapon strength per tier; 1.5 seems it might work nicely.  Then we get:
Everything at tier 3 or below will stay how it is.
Chaos gun increases to 9.
Silksteel increases to 6.
Fusion laser increases to 14.
Photon Wall increases to 9.
Tachyon bolt increases to 18, and Plasma Shard to 20.
Quantum laser increases to 27, and Neutronium to 14 defense. 
Graviton gun increases to 40, and antimatter to 20 defense.
Singularity increases to 60, stasis field to 30, and string disrupter to an impressive 75 strength but 100 cost.
Psi attack and defense will increase with the matching weapons/armor, as will resonance weapons/armor and pulse armor.

Then the "minimum weapon based on armor" just has to be capped at 4.

Then with the reactor values from above, a singularity squad costs 60 minerals without abilities; throw on stuff like drop and wave or blink* and it'll be high enough.

The only concern is the rollover issue.  Perhaps the answer is to simply place a strict maximum cost of 100 rows (1000 minerals with +0 INDUSTRY).  It would mean a discount when getting the absolute maximum of everything, but that'll still be extremely expensive.

*Blink costing 2 as it clearly should, that is; it's quite powerful.

Finally, minimum reactor-based cost should probably be changed to compensate; 2Xreactor seems good.

[Nexii]

I assume there's no (easy) way to just prevent designs over X minerals.  A shame.  Btw with a quick test it seems the rollover is at 2560 minerals.  You can design a unit >3k in the game as is and the actual cost will be 2560 less.

I like the idea to make better weapons/armor more decisively beat the older models though. 

What I was getting at with Native life forms, was that unfortunately they would get very cost effective against the more expensive units (since they ignore reactors).  But I suppose as long as the native life forms (and PSI units) don't beat out gun/unarmored Empath/Trance units in effectiveness, then it will be okay.  PSI would get a lot more expensive at Tier 7 so it might work out.

[Yitzi]

I assume there's no (easy) way to just prevent designs over X minerals.

Well, you can always houserule it.

A shame.  Btw with a quick test it seems the rollover is at 2560 minerals.  You can design a unit >3k in the game as is and the actual cost will be 2560 less.

Is that every case, though?  Or are there some cases where it treats it as less?

Although since it's the result of higher-level code rather than direct assembly language manipulation, it should all be the same, based on what type of variable it was initialized as.

I like the idea to make better weapons/armor more decisively beat the older models though. 

It still won't be certain, by any means, but it will be a substantial difference (though a bit less than the very first, i.e. synthmetal, laser, and impact.)

What I was getting at with Native life forms, was that unfortunately they would get very cost effective against the more expensive units (since they ignore reactors).

On the other hand, you can defend against/attack them with cheap units, as you said.

More of a concern would be a mixed force, but that requires including the more expensive units, and them possibly getting killed.  And of course once you have 100 minerals/turn in a base, worms being only 40 or 50 isn't worth that much.

[Nexii]

The only other thing I would add is tweak some of the chassis multipliers, if you keep it (Armor + Speed) as it is.
 
(A + S) factors for all units:
For attack infantry (1 + 1) / 2 = 1 factor (+10 minerals?)
For rovers (1 + 2) = 3 factor
For hovertanks (1 + 3) = 4 factor
For air units (1 + / 4 = 2.25 factor
For foils (1 + 4) / 2 = 2.5 factor
For cruisers (1 + 6) / 2 = 3.5 factor

Now if unarmored is Tier 0, and thus a value of 0, we'd have this instead:
For attack infantry = 1/2 factor (+10 minerals?)
For rovers = 2 factor
For hovertanks = 3 factor
For air units 8 / 4 = 2 factor
For attack foils 4 / 2 = 2 factor
For attack cruisers 6 / 2 = 3 factor

I might take out the flat +10 cost to infantry.  I could see them scaling a bit too well at such a low multiplier though.  Attack infantry do get mowed down by copters...a lot more efficiently than rovers/hovertanks.  I'd argue for a bit lower factors on hovertanks and foils/cruisers.  Maybe a little more cost on copters or put them to flat 8 move and give them 100% damage if they don't return to base.  Maybe if air units didn't also exert ZOC they wouldn't feel as undercosted.  If Air Superiority becomes Cost = 0, I could see more of the 'put a stalling needlejet out there' strategies.  I'm not sure what the cost effective counter would be to that...SAM infantry maybe?

[Yitzi]

Now if unarmored is Tier 0, and thus a value of 0

You misunderstand.  "Tier 0" is just my label for the starting stuff, with the minimum value of 1.  It wouldn't be reduced to a value of 0; such would not be good and might even break the game.

Maybe if air units didn't also exert ZOC they wouldn't feel as undercosted.

Of course, that ZOC comes at the cost of an aircraft (and is fairly ineffective anyway) once the enemy has SAM, as aircraft pretty much never has armor (it gets quite expensive.)

I'm not sure what the cost effective counter would be to that...SAM infantry maybe?

Indeed it is.  SAM would still remain not-free on land units no matter what, but even mixing in a few SAM infantry in your attack force (and remember, land SAM doesn't have a penalty to attack land units) will demolish that tactic.

The only change needed for that is to make SAM available earlier than needlejets, so that there usually isn't a window where ZOC delaying works.  Synthetic Fossil Fuels seems a good place to put it.

[Nexii]

Agreed.  SAM definately should come earlier at Synthetic.  I think this was the window I was getting at where you don't have air but your ground units can't advance.  This is a good argument for air units not having armor (even a 1/3 move rush attacking infantry will beat an aircraft).

Ok so if the min cost is 2R, all units cost other than infantry have a min cost of: 10, 20, 40, 80.  But if you tack 10 on for infantry, the min costs go: 20,30,40,50.  I think infantry should go 10,20,30,40 for minimum cost.  I'm not strictly against min costs, they're good in that they discourage using obsolete weapons/armor on new reactors.

Air units with AS still hard counter rovers/hovers.  Not really sure if this was intended.  I'm fine with normal air units being a hard counter...that's realistic.  Maybe AS air shouldn't be able to hit ground if it's Cost:0.  It's more the AS copters I think would be a bit too good.  They can clean up so much attack aircraft/rovers/hovertanks.  As well they'd be good against air.

[Yitzi]

Agreed.  SAM definately should come earlier at Synthetic.  I think this was the window I was getting at where you don't have air but your ground units can't advance.  This is a good argument for air units not having armor (even a 1/3 move rush attacking infantry will beat an aircraft).

Well, they actually can have armor, it just gets expensive, especially with higher-level reactors.

Ok so if the min cost is 2R, all units cost other than infantry have a min cost of: 10, 20, 40, 80.

20,40,60,80, actually.

I think infantry should go 10,20,30,40 for minimum cost.

I disagree, as scout patrols should be the only 10-cost units.

Air units with AS still hard counter rovers/hovers.

With AS?  Let's take the case of tier 4 with fusion, i.e. a bit after air power but not that much.
The air units have Chaos guns, at 9 attack, reduced to 4.5 for ground strikes.  They cost 9X(1+8)/5/4, rounds to 4 rows, i.e. 40 minerals.
The rovers include some offensive units, but probably also some defensive ones.  (If you send undefended units against an enemy with air power, you deserve what happens.)  The defensive units have Silksteel, for 6 defense, with a cost of 2X(6+2)/4=4, so also 40 minerals.  So the rovers have a slight advantage at the same cost, or a huge advantage at the same cost with AAA.

Now, without ground strike, the aircraft have the advantage without AAA and it's substantially closer with AAA.  If the force is attacker-heavy, it may be possible to use air units to take them out at below cost-effectiveness and then make up the difference by killing the attackers with choppers.  But I don't really see that as a problem, as rovers/hovertanks are good against infantry (as they have greater mobility and can use that to fight in open ground as the attacker, where they'll have the advantage even with ECM), so you get a weak rock-paper-scissors-esque pattern, which is very good for balance.

Maybe AS air shouldn't be able to hit ground if it's Cost:0.

That'd be harder to code, but it's easy (can be done just with alphax) to reduce its strength as much as you want.  But personally, I think that it's ok for rovers to need defensive units mixed in once aircraft becomes available.