Bonus Ship Ommission File

[keith.lamothe]

Ok, you try defending against a big bomber wave with just missile frigates and let me know how that goes for you

Or, for that matter, defending against a big missile frigate wave with just fighters.

Or even defending against a big fighter wave with just bombers, though that may work if you have a few forcefields protecting the station.  I thought that was the case with missile frigates too but the AI reminded me otherwise yesterday.

This is a fallacious argument known as a straw man.

It would be if you were not using such sweeping generalizations, but as it stands I was just providing counterexamples to your statement that "Everything else can be dealt with by 'whatever'".

To make it a little less obvious: try defending against a bomber wave with just bombers, or even against a missile frigate wave with just missile frigates.  Or with anything that has, say, 4x or higher bonuses but (insert hull type of attacker here) isn't one of them  It is really not the same as if you use the counter.  And there are several fleet ship types (more than can be shuffled off as "exceptions") that are not countered by the bomber where you really want them dead now, and if you just try to deal with them with "whatever" without making sure you've got enough counter in there, you're going to get torn up pretty bad.

My point was, that with any blob of ships that I own, if I have bombers, I beat up everything.  If I don't have bombers, I beat up everything...except big things.

It doesn't have to be bombers, there are other ships that fill a similar role against the heavy stuff, but if you don't have any of that around, sure: if you have a fleet that lacks a bonus against Heavy/UltraHeavy/Structural, there will be important things you will not kill efficiently.  Is that a problem?

[Wingflier]

1. Even if two ships have the same (or similar) health, armor, armor peircing, DPS, damage per shot, cap, etc, with multipliers, we can change how they interact with other units. There are a couple ships out there that are mostly the same except for what they get a bonus against. Now, it could be argue that such "clones in generic stats, but different in match-up based stats" is bad for a game. In which case, these type of "almost clones" would need to be differentiated somehow, or just merged into one ship type (with 100+ fleet ship types, it's not like its going to be missed )

I think that if two ships have almost exactly the same stats, and just different multipliers, these ships need to be different somehow, or like you said, one just needs to be removed from the game.

2. In the example of Sentinel Frigates having low DPS, high armor peircing, given the way armor currently works, this means it would get low DPS against everything. Aka, it would get just as low a DPS against highly armored stuff as it would lightly armored stuff. Hull type multipliers allow for increasing damage to intended types of targets without increasing damage to other stuff as well.

There was an assumption that if were to switch to an entirely Armor-based system, armor would become extremely effective compared to what it is now.  In other words, without some form of armor piercing, you'd basically do abysmal damage to heavily armored ships.  I'm sorry for not clarifying that.

EDIT: BTW, are you arguing that even the current, higher level of management of ship-type vs. ship-type matchups is unscaleable from a "balance maintenance" perspective? (It used to be defined on a per-shiptype vs per-shiptype basis, which was for certain unscalable)

I'm arguing that with the current "multiplier-based" system, each ship's role basically boils down to how well it can kill its multipliers.  If it has no multipliers, it's how well it can kill anything.  Since there are only a limited number of hull types, it stands to reason that a small group of ships will always be the best ships in the game.  I think we have already seen that this is true.

[TechSY730]

Got this in a bit too late, but one more thing:


3. The current system provides a way to differentiate small "physical size" units from large "physical size" units. Sure, it sort of mixes this up with overall durability, but it still provides this system. I know that Starcraft allows some units' weapons to get a bonus or penalty against certain "size classes" (though it is a additive bonus, not a multiplicative one) (and Starcraft II extends this to allow such bonuses and/or penalties to unit properties, like biological, mechanical, and such). This allows for some way to fight "large" stuff, and to keep "large stuff" from just curb-stomping the entire game (assuming you can last long enough to build it ), and to keep "small stuff" from becoming completely useless once the "big stuff" starts showing up. Fighting this trend is one of the reasons for the multplier system, but as Starcraft (and Starcraft II) shows, it is not the only way. (Well, that, and the "phases" and average game pace in Starcraft are far faster, and the economy much more limiting, meaning its harder to get out large stuff in the first place, unlike AI War, where having a cap of all of the build-able "pretty large" stuff by mid game is common).
Maybe a similar system could work for AI war? Some sort of differentiation with far more rigorously defined classifications, and fewer classifications overall, and bonuses/penalties that, with the exception of immunities, do NOT overshadow the base damage?

[Volatar]

One advantage of the system now is how absolutely straightforward and intuitive it is. These proposed changes make it harder to understand the strength of something against other things without doing math in your head on armor and armor piercing and such.

[Wingflier]

3. The current system provides a way to differentiate small "physical size" units from large "physical size" units. Sure, it sort of mixes this up with overall durability, but it still provides this system. I know that Starcraft allows some units' weapons to get a bonus or penalty against certain "size classes" (though it is a additive bonus, not a multiplicative one) (and Starcraft II extends this to allow such bonuses and/or penalties to unit properties, like biological, mechanical, and such).
Maybe a similar system could work for AI war? Some sort of differentiation with far more rigorously defined classifications, and fewer classifications overall, and bonuses/penalties that, with the exception of immunities, do NOT overshadow the base damage?

Starcraft uses basically a "light" "medium" "heavy" system, similar to the armor piercing>armor>dps system I'm proposing.

Basically, units with the "armored" tag can usually survive units without armor piercing (think Helions vs. Siege Tanks).  However, heavy units typically fire slowly, or have some kind of aoe attack, making them weaker to "medium" units (think Marauders vs. Siege Tanks).  Though this is all done with the "armored" or "light" tag, the external attributes of a unit define with it is a medium unit, and therefore strong against "heavy" units.  Light units then counter Medium units (think Marines vs. Marauders).

I still think it's distilled down into 3 basic categories, with the most important classifications coming from a units external attributes (such as flying, cloaking, siege mode, etc.), similar to what I'm suggesting.  I think with Starcraft's system you can focus on what makes each unit unique, instead of focusing on what it's designed to kill.

One advantage of the system now is how absolutely straightforward and intuitive it is. These proposed changes make it harder to understand the strength of something against other things without doing math in your head on armor and armor piercing and such.

I think the new system would be easier, because you can typically tell the units attributes just by looking at it:

"Oh that's a small fast unit, it must have no armor."  "Oh that's a huge, slow, expensive unit, it must have a ton of armor."

In this way the game actually becomes much more intuitive to players, simply by looking at a ships outward characteristics you already know a lot about it, instead of having to memorize endless hull type tables.

[TechSY730]

I feel that, in addition to revisiting distribution of hull types, hull type multipliers should, on average, not dominate base damage by quite such a large margin. Aka, anything beyond 3x (or maybe 4x) and below .25x should be very, very rare (and even 3x (or maybe 4x) should be rare, like only for highly specialized units, like the acid spitter, and things along the 1.25 and 1.5 should be much more common for ships that have multipliers). This would require some serious balance to keep the game from slowing down to a crawl, though (in particular, base damages would need to be reviewed)

[Draco18s]

Try defending against a bomber wave with just bombers, or even against a missile frigate wave with just missile frigates.

Aside from the fact that it never happens?

I'd happily throw "just bombers" at a bomber wave...because of the turrets I put down.  My bombers absorb shots from the AI units quite well.  It's a war of attrition, but...I've got 40 turrets the AI doesn't have beating down on their side.

Add tractor turrets and I get an extra layer of damage screening (bombers SUCK at damaging tractor beam turrets )

Voila, not only do I win, I win without even thinking about it.

[Kahuna]

I feel that, in addition to revisiting distribution of hull types, hull type multipliers should, on average, not dominate base damage by quite such a large margin. Aka, anything beyond 3x (or maybe 4x) (or below .25x) should be very, very rare (and even 3x (or maybe 4x) should be rare, like only for specialized units, like the acid spitter, and things along the 1.25 and 1.5 should be much more common for ships that have multipliers). This would require some serious balance to keep the game from slowing down to a crawl, though (in particular, base damages would need to be reviewed)

Boom! Problem solved!

[TechSY730]

3. The current system provides a way to differentiate small "physical size" units from large "physical size" units. Sure, it sort of mixes this up with overall durability, but it still provides this system. I know that Starcraft allows some units' weapons to get a bonus or penalty against certain "size classes" (though it is a additive bonus, not a multiplicative one) (and Starcraft II extends this to allow such bonuses and/or penalties to unit properties, like biological, mechanical, and such).
Maybe a similar system could work for AI war? Some sort of differentiation with far more rigorously defined classifications, and fewer classifications overall, and bonuses/penalties that, with the exception of immunities, do NOT overshadow the base damage?

Starcraft uses basically a "light" "medium" "heavy" system, similar to the armor piercing>armor>dps system I'm proposing.

Basically, units with the "armored" tag can usually survive units without armor piercing (think Helions vs. Siege Tanks).  However, heavy units typically fire slowly, or have some kind of aoe attack, making them weaker to "medium" units (think Marauders vs. Siege Tanks).  Though this is all done with the "armored" or "light" tag, the external attributes of a unit define with it is a medium unit, and therefore strong against "heavy" units.  Light units then counter Medium units (think Marines vs. Marauders).

I still think it's distilled down into 3 basic categories, with the most important classifications coming from a units external attributes (such as flying, cloaking, siege mode, etc.), similar to what I'm suggesting.  I think with Starcraft's system you can focus on what makes each unit unique, instead of focusing on what it's designed to kill.

IIRC, there are discrete, "enumed" stats in Starcraft and Starcraft II (its been a while)^*, your observation that those (and "lookup tables" keying off of those) do NOT dominate the "general purpose" stats. I like this system to, where if there is a explicit modifier based on a non-numeric, "is-or-isn't there" property of the target, it would not overwhelm normal damage, but rather be just a nice bonus (something that helps with this is that most "per property" modifiers are additive, not multiplicative, though there are exceptions)
Although I wouldn't necessarily like it quite that extreme in AI war, I do agree that the "general", numeric stats should be the predominant factor in unit effectiveness, with any multipliers there just as a nice small to medium bonus, or to enforce immunities or special cases (like scouts).

^*It is important to note that there is no inherent connection between, say, size classification and numeric armor in Starcraft. You could have a small unit with 8 armor, or a large unit with 2 armor. However, in Starcraft, in most cases, that armor plays a bigger role in most matchups than the size matchups do. Aka, the "classes" of normal, standard offense units formed in Starcraft are very weakly influenced by their "enumed" properties.


Also, it is important to note that these "enum" and "enum keyed" properties being less dominant only applies to normal firing damage. Offensive special, active abilities are frequently keyed very strongly to these "enumed" properties, which gives depth to the stratagy (in some cases, you can sort-of bypass the "general case toughness" with a well placed active special ability with an effectiveness strongly bolstered by a specific property of that unit with the "general case toughness"). We don't really have this way to "balance out" this general vs. enumed property usefulness, as special offensive abilities are rare in AI war. This is why I don't think going quite as far as Starcraft in "general stats" dominating "enumed stats" is a good idea, as we don't have this other mechanic to really work with much.
Though this could be solved in part by increasing the prevalence of special passive offensive abilities (more units with odd damage calculations, increasing the role of armor which would increase the role of armor rotting, etc.)

[keith.lamothe]

I'd happily throw "just bombers" at a bomber wave...because of the turrets I put down.  My bombers absorb shots from the AI units quite well.  It's a war of attrition, but...I've got 40 turrets the AI doesn't have beating down on their side.

And, dun dun dun! You win, because you had more than just bombers!  Had you not put something down to deal with something that wasn't a big hulking pile of HPs, it may not have gone so well.  Had you just relied on "whatever" to take care of the bombers, without making sure some anti-polycrystal was in there, it may not have gone so well.

Voila, not only do I win, I win without even thinking about it.

If you do your strategic thinking right, tactically you may not have to think at all.  Tactics is largely about winning the battles that weren't already won before they started.  But if you put see a huge bomber wave incoming you have to think about having enough laser turrets and other anti-polycrystal stuff.  And of course the non-DPS stuff like tractors and gravs and FFs, both in terms of how many to deploy and where to put them, etc.  If you try to do it without thinking about any of that stuff (mainly, ahead of time), a big bomber wave is going to do an excellent imitation of a velociraptor

[TechSY730]

I feel that, in addition to revisiting distribution of hull types, hull type multipliers should, on average, not dominate base damage by quite such a large margin. Aka, anything beyond 3x (or maybe 4x) and below .25x should be very, very rare (and even 3x (or maybe 4x) should be rare, like only for highly specialized units, like the acid spitter, and things along the 1.25 and 1.5 should be much more common for ships that have multipliers). This would require some serious balance to keep the game from slowing down to a crawl, though (in particular, base damages would need to be reviewed) [UPDATED QUOTE]

Boom! Problem solved!

More like, several weeks or months of painstaking, tedious balancing effort, with several of the betas in-between with painfully, nearly unplayably bad balance (both because the new balance model has only been partially applied, and once all units get reviewed, before a "second pass" can be done to see if any tweaks will inevitably be needed, or oversights corrected) sometimes spread out, sometimes several in a row; intense debates about the new roles of ships that were made more or less duplicates of another ship because of the reduced importance of "enum-stat" based properties; figuring out a new average armor (or new armor mechanic) that would allow this new balance target to be reasonable; trying to adapt the meta-game to the new system; figuring out how a unit being AI only or human only would influence what it's balance target should be in this new model; and many other things. Problem solved!


This is not just a hypothetical situation; I have been around for 1.5 rounds of these sorts of cycles. (The 4.0 - 5.0 wasn't as painful as the last SlimDX version - 4.0 cycle, which is why I am counting it only half). However, I do feel like it will be beneficial to the game overall, but it will be a painful transition in the meantime.

OK, granted, the biggest aspects of the game this would change is overall "longevity" and "durability" of units, and how units are informally "classified" due to a shift in relative importance of stats. The last SlimDX version - 4.0 cycle was even more painful, as it shifted these aspects and many more.

[KDR_11k]

There are a few key hulls but many others feel rather arbitrary. That makes it hard to remember what's dangerous and what isn't. The high multipliers that are going around (the often cited Starcraft has multipliers range from 0.5x to 2x) mean that running into the wrong enemy can instakill a part of your fleet. But those enemies aren't rare, high threat targets, they're the regular enemies that are mixed into every encounter. A guardian gets special attention from a player but something like e.g. a bunch of missile frigates waiting for your spire starship is easy to miss. It's like having every third infantryman in C&C equipped with an instakill RPG (actual peak damage in C&C is fairly low, even against preferred targets you usually need maybe 5 shots with heavy weapons and a few secs with light ones, it just takes FOREVER when you use the wrong weapon instead). The multipliers in AI War have reached the status of "here's a list of what you can bother using this unit against, if it's anything else just forget this unit exists". Some of them should be renamed for clarity (e.g. Structural should become Force Field), maybe hull types should be replaced by unit roles (e.g. "this does triple damage to raiders") but that would perhaps reduce the variety of the unit types (but then again would you ever try to have more than one fleet ship of a type? And if you do would you actually split them up or just send both types?).

Of course the flip side is that we NEED this high lethality if we want to have a chance at stopping a pile of AI units blindly charging towards our bases because there's no terrain to stop them, they can move through our forces, building enough tractors would eat your cap in no time and gravs must be unlocked first. So it has to be possible to gun down the vast majority of a wave before it gets in range of the command station. That's basically balancing for tower defense with no maze and 1k+ creeps per  larger wave.

Using no damage multipliers is possible, e.g. Zero-K has an irrational hatred of armor types, but that doesn't mean there are no counters. Other games can rely on their combat physics to produce certain outcomes (e.g. in Company of Heroes a tank shell will instakill an infantryman but the chance to hit something that small is already low and with cover it becomes near zero, other games may have slow artillery shells that the raider units will just run away from) but AI War has no combat physics, things deal damage and that's it.

[keith.lamothe]

On the multipliers being high: in 3.0 some of the multipliers were as high as 100 (inflitrator vs fortress; based on ship type, not hull type), so it's been a long road to the current situation of multipliers only very rarely getting to 10x, and generally being between 2 and 5.

Even with them that high I've heard complaints about them being too low and that leading to a feeling that the units lack diversification.

My personal preference is for systems with no multipliers (except maybe "good vs shields", "good vs armor", etc), and just relying on combat physics like how fast a specific gun mount can track a target moving across its field of fire, the concept of field of fire at all, etc... but that doesn't fit the scale of this game, either at the "can the CPU/RAM be expected to simulate/contain all that" level or at the "can the player be expected to manage all that" level.

[Hearteater]

Since there was a little veering into armor, and since a good armor system will have an effect on what the multipliers can look like, I'll throw out a quick system that would make it easy to have ships specializing in armor piercing be different from those that don't, while also being both simple to both understand and implement:

Rate unit's Armor as Mark I - Mark V (with some ships have no armor, effectively Mark 0).  Armor reduces all incoming damage by a percent as follows: Mark 0: -0%, Mark I: -20%, Mark II: -40%, Mark III: -60%, Mark IV: -80%, Mark V: -90%.

Armor Piercing is also rated from Mark I - Mark V (with some ships having no armor piercing, effectively Mark 0).  If the attacker has an equal or greater Mark Armor Piercing than the target's Armor, then the armor is completely ignored.  Otherwise the armor is fully effective.  This means you have distinct matchups between ships.  Mark V armor ships really want Mark V AP ships to counter them.

Armor Rotting reduces the Mark of Armor by one (and doesn't stack within a single ship type, but does stack between multiple ship types, so Acid Sprayers and Autocannons together could knock off 2 Marks of armor).  This gives those ships the powerful ability, but also makes it very clear intuitively.  Armor can't be reduced below Mark 0, so ships with no armor to start with aren't affected by armor rotting.

Lastly, Armor Boosting adds one Mark of armor, capping at Mark V.

[TechSY730]

My point was, that with any blob of ships that I own, if I have bombers, I beat up everything.  If I don't have bombers, I beat up everything...except big things.

It doesn't have to be bombers, there are other ships that fill a similar role against the heavy stuff, but if you don't have any of that around, sure: if you have a fleet that lacks a bonus against Heavy/UltraHeavy/Structural, there will be important things you will not kill efficiently.  Is that a problem?

No. The complaint is that similar sorts of arguments can't be made for many of the other common hull types, like light, primarily because there aren't nearly as many important things in other hull type categories, and overall, by far the most durable stuff is in the heavy, ultra-heavy, and structural hull types. Aka, BOTH important and durable targets are strongly disproportionate spread in these hull types. If your fleet-ball lacks something with a, say medium hull type bonus, or composite hull type bonus, you are nowhere near as bad off in terms of what your fleet can accomplish in terms of what will progress your game. Sure you may run across something tough that will give you trouble with those hull types you neglected, but the chances of that happening are far less than the chances of you running against something you can't really deal with if your fleet didn't have something with a good heavy, ultra-heavy, and/or structural bonus.

Structural I can understand, as how it is handed out is based on other unit properties and unit roles (only immobile stuff get structural, this counts ff-modules as they cannot move on their own accord). However, giving most of the important stuff and highly durable stuff into these two categories makes all the other hull-types seem very much less worthwhile.

Basically, it feels wrong that some non-special purpose hull types (scout, command-grade, and structural are what I consider special-purpose) are so far and beyond more important than other due to the distribution of things in those types, even if it is possible for a system to be balanced with this sort of skew of distribution.