Bonus Ship Ommission File

[Draco18s]

Second, what "hull types" are modelling is irrelevant. Size, armor composition, class.  Doesn't matter.  You can call them whatever you want.  They can be colors.  The final result is dividing ships into X categories for receiving damage.  Any argument about what we call those categories is just an argument about aesthetics.

Actually, it does.  Take a unity with 100 hp and give it "light" hull type because it has less than 1000 hp.

Now create a unit that does 500 DPS with a 2 multiplier vs. "light."

Does that multiplier even matter?  No, it doesn't.  The first unit still dies in 1 hit.

Now let's look at "heavy" type which is anything with more than 5000 hp.  We'll give this unit 6000 hp.

Same attacking unit, same 500 DPS.  Takes 12 shots.  We decide that "12 shots is too many" and give it a x2 multiplier vs. heavy.

Now it dies in 6.

That's the problem.  Because hull types are divided by how beefy they are, everything ends up shifted towards 1 hp.  If it already had 1 hp, nothing changes.  And things with millions upon millions of hit points are classified as "heavy" requiring a lot of things to have multipliers vs. heavy just to kill them at all, unlike light and ultra-light where if you look at them funny they explode.

Only when a unit dies too quickly it's hp is increased, rather than downshifting the multipliers on other untis (because then all the other units with the hull type need to have their HP decreased).

Back in version 2, things had "thousands" of hitpoints.  In 3 it got upped to "tens of thousands."  In 4 it was upped again to "hundreds of thousands."  In 5 and 6 it's now up at "millions."  Because every single-unit balance has either tweaked HP upwards (so it dies less fast) or its multipliers upwards (so it kills things faster) or its damage listing upwards (leaving post-multiplier DPS the same, but increasing DPS vs. non-preferred targets).

That's the other reason Armor is so worthless: it was originally meant for units that had "tens of thousands" to "hundreds of thousands" of hitpoints.  Not millions.

So (at least) three things need to happen:

1) Make hull types a material which is determined independently from its hit points.  "Polycrystal" is a great example.
2) Fix armor values being meaningless--this is where the current "hull sizes/types" can be used.  Heavy?  More armor.  Light?  Less armor.
3) DPS and hitpoints need to be revisited so that preferred targets aren't blown away like dust in a hurricane.

[TechSY730]

(a bunch of stuff)

Thanks, this is what I was sort of trying to communicate, but you did it in a clearer fashion. (And even provided some nifty concrete examples demonstrating the pitfalls of assigning hull types by overall durability given by other stats)

There is another option for hull type names/concepts, ship role. One example is artillery hull. Although that is a bit dangerous as well (many ship roles have "expected relative durabilities", even if it isn't as strong a connection as the "durability names" hulltypes), but it can be nifty as well (for example, polycrystal in large part is the "bomber or bomber-like" hull type). The key is that both of these things to model do not inherently tie to overall durability, which ultra-light, light, medium, heavy, and ultra-heavy do.
However, for consistency, I say that hull types should have a common naming scheme. And given the choice between ship role and hull material, I would choose material, as it is less likely to conceptually "overlap" what should be modeled in other stats.

[Wanderer]

1) Make hull types a material which is determined independently from its hit points.  "Polycrystal" is a great example.
2) Fix armor values being meaningless--this is where the current "hull sizes/types" can be used.  Heavy?  More armor.  Light?  Less armor.
3) DPS and hitpoints need to be revisited so that preferred targets aren't blown away like dust in a hurricane.

First, let me state that in general I agree with most of what is written above.  However, I also realize the massive undertaking a complete game rebalance would take.

There is something else that should be pointed out here, however.  Keith has taken great pains to make sure that it's NOT ship on ship balancing, but shipCAP vs. shipCAP balancing.  The problem with cap balancing however is that it doesn't take Lanchester's into account, that being the decay of firepower over time when small units are facing larger ones.  If you're not familiar, think APCs vs. Infantry on a battlefield.  While the Infantry will eventually swarm the APCs, the damage any one of those units inflicts is so high before the infantry can land the killshot that there's a significant firepower decay.  In a way, the mulitpliers for the little guys are supposed to help balance off the Lanchester's laws by upping their firepower before decay reaches too high. 

Selective engagements however were tougher prior to 'preferred reinforcing' by the AI.  Now you're more likely to bring in a selective strike team against a planet that's 90% of three ships instead of always needing the full fleet-ball for a generic on generic brawl.  Guerilla fights are actually more likely with this new mechanic, but I digress.

An end to end rebalance will be, at best, difficult.  How do you propose to balance a 4/8 ship cap against one with 96?  How do you make it so that it's not just a game of who's got the biggest fleet?  There's a lot of reasons for some of the imbalances, and they make for a very interesting game.  There's only a handful of ships that you can point at and go: "God ship, gimme."  BladeSpawners, Snipers, and Zenith Snipers for their range.  TDLs for their pathing control and range.  The new Spire Railclusters, (now, only high end) Maws, and Zenith Beam Frigates for their amazing small ship crowd control.  SSBs as fleet-level starships with cloak/raid abilities, and sometimes EyeBots because you just REALLY need to get under a forcefield... though Raid SSs also fill that role.  Youngling Hordes for incredibly cheap swarming assaults (though, Dark Spire REALLY counters that heavily now).

Note, none of these are unbalanced based on their hit points nor on their multipliers, but because of their specialties.  I agree the armor mechanic itself has been made nearly moot at this point, but I'm okay with that until Keith and the community can revisit armor and make it make sense again.  Until then, I'd prefer to see it as an ignored stat... like charisma on the barbarian.

[Faulty Logic]

There is a counter to Lanchester's law: overkill. Very often, the last shot wastes some of its firepower, and more targets means more wastage, pushing it back toward the swarm.

Hull types: I think it could add some to the game, but its game improvement/time invested would be far inferior to many other avenues.

[Draco18s]

First, let me state that in general I agree with most of what is written above.  However, I also realize the massive undertaking a complete game rebalance would take.

Keith has taken great pains to make sure that it's NOT ship on ship balancing, but shipCAP vs. shipCAP balancing.

Oh sure.  It'd be hard, and sure, Keith has gone to great pains doing it on a cap-to-cap basis.

But there are still flaws in the result.

How do you propose to balance a 4/8 ship cap against one with 96?

It's hard, I'll grant you that.  But it's not impossible.

Will it take a lot of work?  Sure.

[TechSY730]

Yea, addressing these sorts of things will take quite a bit work, time, math, and deep thinking. Thus, I don't expect all this rebalancing to be pushed out in a patch later this week.

Also, the interaction between cap count, cap strength, individual strength, mutlipliers, weapon range, and movement speed and many other factors are really, really difficult to consider, thus making units that have one of these"extremes" even more tricky to keep useful or from being OP, as well as keeping the units with an average of these useful but not OP.
Many RTSs try to avoid this issue by just reducing the distance between the min and max of all of these various stats, plus not having as many build-able unit types, thus reducing how much failing to consider all aspects can hurt balance. I'm glad AI War did not go this route, but seeing how complicated it gets to model how all these stats interact with each other, even at a rough level, I can see why many other RTS makers avoid getting into situations where they would have to.

Hull types: I think it could add some to the game, but its game improvement/time invested would be far inferior to many other avenues.

Are you suggesting that we just leave hull types as is for now, as there are bigger "fish to fry", or that hull types should be just removed?

[Faulty Logic]

Are you suggesting that we just leave hull types as is for now, as there are bigger "fish to fry", or that hull types should be just removed?

The former. Hull types (and manuevering so that you bring your counter against the AI, while preventing it from doing the same) are a big part of the tactics of this game.

For the moment, I think refinements to champion balance, hacking improvements, more interesting hybrid stuff, some AI behaviours that are still cheesable, buffing 10/10, buffing the AI Homeworlds, Spirecraft balance modifications, Golem balance modifications, and command station balance, while not necessarily "bigger" fish, are more efficient fish.

And I don't mean a complete freeze on fleetship balance changes, just not a major overhaul.

[Draco18s]

The other thing to consider is "finickiness."  That is, how much per-unit micromanagement is needed to utilize their strength.  E.g. player-use vs. AI-use.

Some units (like the Etherjet tractor, which I like to harp on) heavily favor the AI.

Other units (blanking on an example) heavily favor the player.

And that factor has to be considered, and there are various ways to tweak units for AI use: buffed/nerfed "cap" or tweaked unit stats.

[Wingflier]

One of the reasons why I felt that the ultra-light, light, medium, heavy, and ultra heavy hull types were, well, a mistake (an argument could be made for the swarmer hull type as well). They sort of confuse multiple systems. Overall durability already has some good stats for it, armor and HP. There is no reason to "stack the books" against the lighter stuff even more by tacking on a hull type for small stuff with it. In my mind, hull type should be just that, a type, aka, a material. There is no need to artificially cheapen or make more valuable a hull type multiplier by having some hull types be tied, (de facto even though it isn't in code) to a concept of overall durability. We already have stats that model that.

I definitely agree with this.  Giving a "light" hull-type to a "light" unit is simply a bit superfluous, and giving extra bonuses to that unit based on the fact that it's light, is simply contradictory design.  Light units don't need huge bonuses against them...they're light.

Hmm, from what I have seen, most other RTSs don't have many hull-type multipliers, if any. Instead, they seem to group multipliers based on unit physical size (small, medium, large), and let a numeric armor stat model overall armor durability. Hull materials, if mentioned at all, are pretty much just "fluff". (Like Starcraft and Starcraft II)

Agreed.

I've talked about this at length before, but basically most of what we accomplish with "hull types" now could be better and more untuitively addressed using a proper armor and armor piercing system.

You'd basically have 3 class of units:

1. Units with high base DPS.
2. Units with high armor.
3. Units with high armor piercing.

And of course any hybrids of those 3 categories.  Armor counters units with high DPS.  High DPS counters units with armor piercing.  Armor piercing counters units with armor.

What we know now as "light" unit would most typically be high DPS units (they already are), decent against most things without high armor.  Their weakness is that they are still relatively light, meaning they die quickly to aoe and other light units (like Fighters in a Sci-Fi Universe).  People get freaked out when I mention this concept, fearing "light" units will become fragile glass cannons...they already are lol.  The only problem now is that they typically underperform compared to their heavier cousins, despite being so easy to kill.  There's no beneficial trade-off.

What we know now as "heavy" units would be units with high armor like bombers and their counterparts, as well as things like Hybrids, Golems, etc.  They don't do as much damage as light units (for their cost), but they make up for it with their survability, and they are especially good against heavier targets.

"Medium" units would be somewhere inbetween "light" and "heavy", having some combination of high dps, armor, or armor piercing, and many units could fall into this category such as frigates, autocannon minipods, spire armor rotters, and several other types of bonus ships that aren't very useful at the moment.

Theoretically, we could balance the entire game without using hull types at all.

Think about it.  If what makes a Raider powerful is that it can quickly get behind enemy lines and deal damage to important targets, then why do you need hull types for that?  A raider is a quick "light" unit with high DPS and good armor piercing; a hybrid if you will.  This makes it powerful against most buildings and structures (including Force Fields, which would now have a high armor value), but extremely weak in a fight because of its low HP.  Suddenly raiders are much more effective than they were before, and we didn't have to do it by buffing their hp or changing their multipliers.

If what makes a Spire Armor Rotter powerful is that it rots armor, then all you have to do is make armor useful for them to have a purpose. 

In other words, we could keep all of the same bonus ships, and by removing hull types, and making a more robust armor system, I don't see how ANY of them would lose their roles, yet many of them would have a more useful role than before.

[Hearteater]

Actually, it does.

No, it doesn't.  You are arbitrarily assigning low HP units to Light hull.  Armored Golems could be assigned "Light" hull type.  Light, Heavy and whatever hull types have no bearing on anything else.  Since this appears a confusing concept, I recommend using the following place holder hull types: infrared, red, orange, yellow, green, blue, indigo, violet, ultraviolet.  That should eliminate the natural bias the current names are being assigned.  As the final step, they can be assigned real names.

So (at least) three things need to happen:
1) Make hull types a material which is determined independently from its hit points.  "Polycrystal" is a great example.
2) Fix armor values being meaningless--this is where the current "hull sizes/types" can be used.  Heavy?  More armor.  Light?  Less armor.
3) DPS and hitpoints need to be revisited so that preferred targets aren't blown away like dust in a hurricane.

1) They already are, you'll notice several Guardians just got their hull types changed, something not possible if hull type was based on hp.
2) The armor system is already under review in a separate thread.
3) This would be necessary after the armor change anyway.

[Draco18s]

while not necessarily "bigger" fish, are more efficient fish.

And I don't mean a complete freeze on fleetship balance changes, just not a major overhaul.

The problem is that there will always be "more efficient fish."  At some point someone has to say "ok, we're doing the overhaul" and buckle down and do it.  Until it's done it's just going to loom there over our heads.  And occasionally, drive people away from the game.

I seriously cannot get anyone I know to play the game (and hell, I don't even have strong feelings for it myself).

And none of them will even consider coming back until that rebalance is done.

No, it doesn't.  You are arbitrarily assigning low HP units to Light hull.

That's the way it already is!

There are a few exceptions (notably that the raid starship line is ultra-light) but for the most part "heavy" and "ultra-heavy" are dictated by how many hit points it has.  Keith has even said as much, numerous times.

[Volatar]

Removing hull types would remove all the specific counters we have now which is such a great part of the game. Removing and counters would move us back towards fleet balling at all times.

[TechSY730]

And that factor has to be considered, and there are various ways to tweak units for AI use: buffed/nerfed "cap" or tweaked unit stats.

There are also ways to adjust how much the AI has to "pay" to get a cap without changing how much the player gets or how much the player pays for it (there is a per-shiptype multiplier for how much the AI can get in a wave given the same amount of "points" to distribute, and I think, but I am not sure, there is a similar multiplier for defensive spawns as well, to deal with the fact that the AI has different goals and a different "micro-ability" thus shifting the balance of certain types of ships.)

[TechSY730]

No, it doesn't.  You are arbitrarily assigning low HP units to Light hull.  Armored Golems could be assigned "Light" hull type.  Light, Heavy and whatever hull types have no bearing on anything else.  Since this appears a confusing concept, I recommend using the following place holder hull types: infrared, red, orange, yellow, green, blue, indigo, violet, ultraviolet.  That should eliminate the natural bias the current names are being assigned.  As the final step, they can be assigned real names.

That is the other thing I was trying to communicate. The names of ultra-light, light, medium, etc give a conceptual connection to HP and durability, even though there is no inherent connection to it. This lead to the "temptation" to assign them as such. Another reason why I think that hull types should be names after materials (or after colors or some such until we can come up with new "sci-fi-ish" enough sounding names for the names that would be replaced), there is much less conceptual bias between material types and overall durability.

EDIT: So basically, you are both right. The NAMES do not impact anything. However, proper DISTRIBUTION is very influenced by what is being modeled, and the conceptual bias that can come from certain names can make it harder to keep in mind that proper distribution.

[Draco18s]

There are also ways to adjust how much the AI has to "pay" to get a cap without changing how much the player gets or how much the player pays for it

What I mean is, it's another factor to consider when rebalancing, and that there are mutliple methods to make the AI "pay" more.  E.g. there's the "usefulness in waves" size multiplier--of which only the bomber has a unique value (0.8 vs. everything else at 1).