1.8 - Which Combat Tests should a Balanced Combat System Pass?

[Larknok1]

User terrycloth had the excellent idea of -- rather than speculate about possible numbers -- provide a guideline of what sort of combat tests a balanced combat system should pass.

This struck me immediately as a brilliant idea -- providing a straightforward way of testing any game-state as balanced.

If a test is currently passed (1.7), it will have a "O" before its number. If it currently fails (1.7), it will have an "X" before its number.

Here are the various combat tests a balanced combat system should pass:

---

Ship tests:

X 1) Corvette torpedo stacks should mostly beat all-medium heavy shield Cruiser stacks, as well as most Battleship stacks.

X 2) Anti-Corvette Destroyer stacks should mostly beat Corvette stacks, and do so with greater survivability than an anti-Corvette Cruiser stack against that same Corvette stack.

X 3) Point-defense (PD) Destroyer stacks should mostly out-compete PD Cruiser stacks at destroying missiles.

O 4) Cruiser (non-PD) stacks should mostly beat Destroyer stacks.

X 5) Battleship stacks should mostly beat Cruiser stacks.

---

Weapon tests:

X 6) PD stacks (either Destroyers or Cruisers) combined with non-PD ships should mostly out-compete pure non-PD stacks against missile/strikecraft stacks.

X 7) Missiles/Strikecraft should mostly beat non-PD stacks of the same ship class.

O 8) Tachyons should mostly out-perform Gigas against heavy armor Cruisers/Battleships.

X 9) Gigas should mostly out-perform Tachyons against heavy shield Cruisers/Battleships.

X 10) Anti-Shield (AS) weapons should mostly out-perform non-AS weapons against heavy shield stacks.

O 11) Anti-Armor (AA) weapons should mostly out-perform non-AA weapons against heavy armor stacks.

X 12) Autocannons should out-perform Basic Kinetics against Corvette stacks.

X 13) Lasers should be better against something than Plasma, and Plasma should be better against something than Lasers.

X 14) Torpedo stacks should out-perform non-torpedo stacks against Cruisers/Battleships.

X 15) Whirlwind missiles should mostly out-perform torpedo / base missile stacks against PD-heavy stacks.

X 16) Lasers / Kinetics / Missiles should all have identical (or near-identical) combat results in the very early game when Corvette stacks fight other Corvette stacks.

---

Current state of balance (1.7): 13 failed tests, 3 passed tests.

---

Any tests that I missed? Let me know!

 

[Zer0k]

I guess the stack vs stack implicitly means they both have the same power? Or is it by fleet usage (as in 8 corvettes for every 4 destroyers for every 2 cruisers for every 1 battleship)?

 

[GloatingSwine]

I guess the stack vs stack implicitly means they both have the same power? Or is it by fleet usage (as in 8 corvettes for every 4 destroyers for every 2 cruisers for every 1 battleship)?

Those are largely irrelevant.

The main real balance factors are either mineral cost, upkeep, or build time because those are the real limiting factors in play.


Really though before you can determine any tests you have to decide what the intended function of each component is within the game.

 

[Draconaes]

Power calculation is kinda wonky. I assumed stack vs stack meant for the same mineral cost, but maybe there's other ways to measure it meaningfully?

 

[GloatingSwine]

3) Point-defense (PD) Destroyer stacks should always out-compete PD Cruiser stacks at destroying missiles.

This is difficult to isolate from other factors like cruiser armour. A cruiser stack will defeat a missile armed opponent better than a destroyer stack because the cruisers can tank more missiles, irrespective of point defence.

To actually test it you'd need to do something like have a cruiser stack with PD destroyer support and a cruiser stack with PD on the cruisers both against a torpedo armed corvette fleet and see which takes less losses. (And the cruiser stack with PD on the cruisers would be better because +fire rate will be better than +tracking, even against missiles since the best missile evasion is 40 and PD3 has 30 tracking meaning an effective evasion of 1%)

 

[Zer0k]

mineral cost

I am not sure this a wise way to approach the balance part. I am not saying mineral cost can't be taken into consideration but I am wary of making it the prime factor.

Power calculation is kinda wonky.

That maybe true, but it still feels relevant enough since it is an indicator of damage and survivability concentration.

You can easily balance based on power then rebalance mineral cost to make it equal across the board.


But to reiterate, I can't quite put my finger on it, but mineral only approach is definitely missing something.

 

[Larknok1]

I guess the stack vs stack implicitly means they both have the same power? Or is it by fleet usage (as in 8 corvettes for every 4 destroyers for every 2 cruisers for every 1 battleship)?

As GloatingSwing said, it's by mineral equivalence.

 

[Larknok1]

There, I finished editing in all the pass/fails to the tests. As you can see, the current state of combat balance is pretty miserable.

 

[Draconaes]

I am not sure this a wise way to approach the balance part. I am not saying mineral cost can't be taken into consideration but I am wary of making it the prime factor.

I'm not expert on balancing, but I certainly feel mineral cost should be a very important factor. I mean, if a battleship fleet can beat a cruiser fleet but costs twice as much, that doesn't tell us much. If a cruiser fleet of twice the size (and thus twice the cost) is able to beat it in turn, why wouldn't a player just build the twice as large cruiser fleet?

If the counter only works when you spend more minerals on it than what it's meant to counter, then it isn't much of a counter.

 

[Hugh Mann Bean]

X 14) Low RoF Anti-Shield weapons (like Kinetic Artillery) should always out-perform High RoF Anti-Shield weapons (like Disruptors) against Corvettes. The opposite should be true against Cruisers/Battleships.

This looks backward. High rate-of-fire weapons should do better against smaller ships, not worse, IMO.

 

[GloatingSwine]

That maybe true, but it still feels relevant enough since it is an indicator of damage and survivability concentration.

It isn't though, because it ignores armour which is almost the single most important factor in survivability, and AP which is the most important factor in dealing damage. Fleet power can be off by a factor of three, it's an almost completely useless measure.

The actual limiting factors on a fleet are: How many you can afford to build, how many you can afford to run, and how many you have time to build.

 

[Larknok1]

This looks backward. High rate-of-fire weapons should do better against smaller ships, not worse, IMO.

You're right -- that was my intention. Lost track of the wording, my bad.

 

[GloatingSwine]

This looks backward. High rate-of-fire weapons should do better against smaller ships, not worse, IMO.

More to the point it doesn't look like a relevant test. Corvettes are not noted for their shield strength but their evasion, this should be testing the specialist anti-corvette weapon (autocannons) against non-specialist weapons (kinetic/laser). Autocannons tend to lose that at the moment btw. Probably due to range problems.

 

[Larknok1]

More to the point it doesn't look like a relevant test. Corvettes are not noted for their shield strength but their evasion, this should be testing the specialist anti-corvette weapon (autocannons) against non-specialist weapons (kinetic/laser). Autocannons tend to lose that at the moment btw. Probably due to range problems.

That's already listed as test 12.

The point is for Disruptors to out-compete Gauss as a meaningful answer to high-shield Corvettes. They don't. But to your point: maybe they shouldn't, and test 14 is irrelevant.

 

[Larknok1]

More to the point it doesn't look like a relevant test. Corvettes are not noted for their shield strength but their evasion, this should be testing the specialist anti-corvette weapon (autocannons) against non-specialist weapons (kinetic/laser). Autocannons tend to lose that at the moment btw. Probably due to range problems.

I've removed test 14 (and renumbered the consequent tests) in accordance with agreeing with you that the previous test 14 was largely pointless.

 

[Dalinski]

Short range / long range testing? The issue being its so easy to engineer a short range battle in Stellaris if you will it.

 

[Larknok1]

Short range / long range testing. The issue being its so easy to engineer a short range battle in Stellaris if you will it.

Interesting. How would you (specifically) phrase a short / long range test? Autocannons should be superior to Base Kinetics if fleets blink on top of one another, but otherwise, Kinetics win? Same for Plasma / Laser?

Type one up and I'll gladly add it.

 

[Dalinski]

Well, I hate to say it, what you are attempting here is in my opinion the single most important thing that could happen to Stellaris.

Across all scenarios does the testing change with short range or long range? Basically a category for each, I dread the extra work it would create you :-(

 

[Zer0k]

I'm not expert on balancing, but I certainly feel mineral cost should be a very important factor. I mean, if a battleship fleet can beat a cruiser fleet but costs twice as much, that doesn't tell us much. If a cruiser fleet of twice the size (and thus twice the cost) is able to beat it in turn, why wouldn't a player just build the twice as large cruiser fleet?

If the counter only works when you spend more minerals on it than what it's meant to counter, then it isn't much of a counter.

It isn't though, because it ignores armour which is almost the single most important factor in survivability, and AP which is the most important factor in dealing damage. Fleet power can be off by a factor of three, it's an almost completely useless measure.

The actual limiting factors on a fleet are: How many you can afford to build, how many you can afford to run, and how many you have time to build.

I am going to combine my answer to you guys because I feel it is one and the same.

First, I need to reiterate, I am not saying we should throw mineral cost out of the window. I am simply saying it making it the sole way of making the tests is flawed as is it doesn't take into account the complete picture.

Mineral equivalent fleets do not take into account the general effectiveness of each fleet, yes it can be a good measurement to see how effective per mineral spent each fleet is, but it doesn't measure how good the fleet is on its own. It will lead to power differential situations where a super effective combo is quite weak on its own outside of the relative strength matchup. This is quite a problem because it seems like we might end up with a disjointed sphere of tests (can't link all tests to each other one way or another). Not to mention, how do you exactly test for this when you have an extreme amount of variability of weapons/defenses: Corvette stacks should always beat all-medium Cruiser stacks, as well as Battleship stacks.
Simply put, a mineral only approach is extremely reductive and might very well do more harm than good if we use it to rebalance.

So perhaps, there should be an accompanying tool in addition to minerals when making balanced changes based on the result of these tests.

Considering the rarity (or difficulty of coming up with them) of these tools, then as flawed as it can be, power can be made useful.

 

[Larknok1]

I am going to combine my answer to you guys because I feel it is one and the same.

First, I need to reiterate, I am not saying we should throw mineral cost out of the window. I am simply saying it making it the sole way of making the tests is flawed as is it doesn't take into account the complete picture.

Mineral equivalent fleets do not take into account the general effectiveness of each fleet, yes it can be a good measurement to see how effective per mineral spent each fleet is, but it doesn't measure how good the fleet is on its own. It will lead to power differential situations where a super effective combo is quite weak on its own outside of the relative strength matchup. This is quite a problem because it seems like we might end up with a disjointed sphere of tests (can't link all tests to each other one way or another). Not to mention, how do you exactly test for this when you have an extreme amount of variability of weapons/defenses: Corvette stacks should always beat all-medium Cruiser stacks, as well as Battleship stacks.
Simply put, a mineral only approach is extremely reductive and might very well do more harm than good if we use it to rebalance.

So perhaps, there should be an accompanying tool in addition to minerals when making balanced changes based on the result of these tests.

Considering the rarity (or difficulty of coming up with them) of these tools, then as flawed as it can be, power can be made useful.

I could only foresee four things mattering:

1) Mineral equivalence

2) Fleet cap

3) Build times

4) Diversity of ship/weapon types

--

Mineral equivalence is clearly important. Fleet cap probably isn't quite as important as you might think -- consider how effective Naked Corvettes were even though fleet cap worked against them.

Build times is super important -- it's part of the reason naked corvettes worked so well.

Diversity of ship/weapon types: it's unfortunate but true that -- outside of really good synergies like Tachyons + Kinetic Artillery -- a mono-culture of one weapon / ship type is almost always better, because targeting AI synergizes.

--

Also: please consider that the tests above are general outlines, and not the absolutes that the language "always" seems to suggest. This is obvious by the fact that some of the tests are in direct contradiction:

Missiles should always beat non-missile, non-PD fleets, and all early-game ships should perform equally.

The tests are meant to represent the vast majority of match-ups, on the conditions that they are not in contradiction with another test.