Current balance of starships in exo waves?

[keith.lamothe]

Erm, to clarify my clarification, I was under the assumption we were trying to get a single RelativeStrength number that would be used for all Exo sources (FS, Golems, Spirecraft) and as you had not posted the RelativeStrength values for the Golem or SPirecraft exo waves I was just throwing those numbers out to start.

FS, Golem, and Spirecraft exos do all use the same numbers (exactly the same, no exceptional multipliers for specific ships, etc), and when we shift it to using relative-strength I want all 3 of those to still be looking at the same numbers as each other.  Their _budgets_ vary, but the cost of their choices should not.

I think I've gone and confused myself. What exact numbers are we trying to compile here?

Personally, I want to:

1) Find values for the "RelativeStrength" column of my table on page 1 of this thread such that all game logic current using the numbers in the "EventAttackCost" column can use the "RelativeStrength" numbers instead, in conjunction with:
2) Find a conversion between the current exo-attack budget sizes and what they should be once they're using the RelativeStrength costs.

[keith.lamothe]

IIRC, the Mothership can also mount modules (much like the Avenger).

It actually doesn't, but I wish it did

[TechSY730]

IIRC, the Mothership can also mount modules (much like the Avenger).

It actually doesn't, but I wish it did

And because they can't now, Spire Superdreadnoughts (final exo-wave push version), tieing mothership with a base HP of 800,000,000, have, on average, a higher durability, which again, doesn't seem right.

Then again, this thread isn't about tweaking the ships, but rather their costs, so I will stop harping on trying to convince you to buff the already scary (even in their current state) motherships, or at least on this thread.

(I thought I posted a mantis with this suggestion, but I guess I didn't. *shrug*)

[Cinth]

Something else to note from my play experience that might be relevant.  The AI likes spire shields. A whole lot.  While on its own, the spire shield isn't much of a problem, but when the exo has a bunch of MK Vs, it can and often does make dealing with the more dangerous ships that much more difficult.

Not to mention they do crazy things when pushed into ff lines or densely populated areas. They don't like to stack at all so when they get crammed together, they start mini-teleporting around. It's annoying and often allows them to get behind my lines.

[Diazo]

In that case, the first step is pretty straight forward.

The golems already use there RelativeStrength value for the Golem/Spirecraft exo-waves right?

Multiply that out and you get

HK I: 5940
HK II: 11880
HK III: 17820
HK IV: 23874
HK V: 47520
Mothership: 47520

For reference, if the HK was actually using the linear Mark progression everything else uses, HK relativestrength should be 7136 * Mk when you average those 5 numbers out.

Exo wave budget gets multiplied by the same amount (0.18127) to convert it from ExoPoints to relative strength.

That establishes the top and bottom of the scale, now it's just a matter of determining where stuff lands as in my opinion at least the Spirecraft and Zenith/Spire SS need to move on that scale.

Almost bedtime here though, I'll keep pondering this tomorrow.

D.

[keith.lamothe]

The golems already use there RelativeStrength value for the Golem/Spirecraft exo-waves right?

No, currently all forms of exo use the exo costs for all the units they pick

Or am I misunderstanding what you mean?

Anyway, I gotta run too.  Thanks for the number crunching, we'll figure something out.

[Diazo]

Okay.

To make sure I'm actually looking at the data correctly, on that chart on the first page you posted, what numbers are actually used in the game for what, and what numbers are random placeholders that are not actually used?

And then what were you looking at seeing?

I've apparently charged off in every direction but the right one so far so let's see if I can turn out some useful numbers this time.

D.

[Toranth]

a)  To do equivalent damage, it takes 4,248 fighters to match the DPS of an H/K.
b)  To do as much damage in a single salvo, it takes 3,187 fighters.
c)  If you had Fighters vs H/K 1-by-1 until the fighters killed the H/K, it would take 24,510 fighters.
d)  If you had a group of fighters vs 1 H/K, it would take a group of 1,093 fighters to get the kill.

I tried crunching some Fighter_equivalent numbers for the rest of the ships, like I did for the H/K Mk I above.
Unfortunately, I think I see part of the problem.
As you look at increasingly more powerful units, the ratio to fighter strength goes up exponentially, while d), the group_of_fighters_equivalent goes up logarithmically.

In other words, if you try to rank ships by damage like a) or c), you end up with lots of very similar ships, followed by a huge spike at the high end (Golems, H/K, etc).
If you try to rank ships by a "How many Fighters is this equivalent to?" method like d), you get some huge differences down at the crappy-ship end (Infiltrator), but it flattens out rapidly, and there is not a lot of difference betwen the high-end ships.

I can understand why these graphs turned out that way, but it doesn't help.  Either one of these makes it difficult to split the ships into Tiers for easy pricing based on a single metric.  To do that, you'd need something that is basically linear across all ships.

[Diazo]

I have not crunched any numbers yet as I'm not sure which ones are valid, but I'm going to take the Exo-points and graph it out, then I'm going to take the RelativeStrength numbers that are actually valid and graph them out, then guesstimate the missing numbers on the RelativeStrength graph to make the shape of the two graphs match.

Then take the ExoPoints of the mothership and the relativestrength of the mothership to get a 'ConversionFactor' and scale all exo-budgets by that conversionfactor and you've got the first run of changes at the point where playtesting is probably necessary.

D.

[Hearteater]

I was pondering setting up a test system and running increasing numbers of each ship type at it to get a practical value, but sadly we can't spawn arbitrary numbers of a ship.  Any chance for a "cmd:spawn LazerGatlingIII 230" cheat that would spawn 230 AI Laser Gatling IIIs on the current system at a random "red name" warp point (aka, one with a warp gate on the other side)?  If there was a cheat toggle to disable the flee behaviour when outnumbered, that would be handy too.

[Winge]

One more thing that should be factored into relative strength:  Immunities.  Something that is Immune to Forcefields, Sniper Shots, Tractor Beams or Gravity (after the lead ship goes down) is a lot more dangerous than something without those immunities.

I'd also be curious to see how people defend from Exos (saves, pics, vids, etc).  I could hold off normal 9/9 AIs indefinitely, even at 100 AIP, but I turn on Golems and Spirecraft at 4 and I get wiped out of the galaxy by the first Exo at < 30 AIP...

[TechSY730]

One more thing that should be factored into relative strength:  Immunities.  Something that is Immune to Forcefields, Sniper Shots, Tractor Beams or Gravity (after the lead ship goes down) is a lot more dangerous than something without those immunities.

I'd also be curious to see how people defend from Exos (saves, pics, vids, etc).  I could hold off normal 9/9 AIs indefinitely, even at 100 AIP, but I turn on Golems and Spirecraft at 4 and I get wiped out of the galaxy by the first Exo at < 30 AIP...

I don't have any graphic aids at the moment, though I can get you a save later this evening.

Basically, on my chokepoints (both of them, I try to use a 2 chokepoint system), I use tractor beam turrets to pull away the fleet ships. Grav turrets pull off groups that don't have a grav immune leader.

Then on my core worlds, I have a black hole generator (to pull off things that don't have black hole immunity, which is most of them), to stall even longer. Also, I have two forcefields (1 Mk. I regular, 1 Mk. I armored) over the wormhole to my home, pulling off the stuff without FF immunity (partial or full). Also, I have an additional set of grav turrets and tractor turrets on both the incoming and to my HW wormholes on my core planets.

Finally, on my HW, yet another detachment of grav turrets and tractors, combined with trader goodies (superfortress especially).

With all this, very little actually makes it onto my homeworld, most of the rest is stalled too long to really make it.

All this is combined with the usual spattering of turrets (on the chokepoints, core worlds, AND homeworld), fortresses, forcefields, etc.
And of course, I have my fleet helping out, splitting them up if needed.



....



Or did you mean early game exo-defense?

[keith.lamothe]

To make sure I'm actually looking at the data correctly, on that chart on the first page you posted, what numbers are actually used in the game for what, and what numbers are random placeholders that are not actually used?

By column:

ObjectType - not used in sim math

ShortName - not used in sim math

EventAttackCost - used only in determining exo composition

RelativeStrength - For the HKs this is currently ignored, for everything else it is used when computing how many can go in waves, reinforcements, special forces spawns, how much they count for when killed near a Dark Spire spawner, etc, etc.  Tons of stuff.  If you're really curious I can put together a complete list of what uses it, but it would be later.

IndividualHealth - just included as reference in the table; it's not used when determining composition of whatever

IndividualMaxDPS - just included as reference in the table; it's not used when determining composition of whatever (though it factors into "Firepower", another metric I'd like to just replace with RelativeStrength, but that's another battle for another day)

And then what were you looking at seeing?

I'm not sure what that sentence means

I would like to find values for "RelativeStrength" and for exo-budgets such that it can be used in every place that "EventAttackCost" is currently used, so that EventAttackCost can be removed entirely.

I do not think many of the RelativeStrength numbers need adjustment for this, though obviously there are the HKs and mothership and some others that are clearly off.  The main adjustment will be to the exo-budgets.

I've apparently charged off in every direction but the right one so far so let's see if I can turn out some useful numbers this time.

Actually I think you've been aiming at the right target, whether you were doing it for the same reasons I was I can't say

Hearteater: there is "warp in the clowns,LazerGatlingIII,100" or whatever, which launches a normal wave of the specified type with a the specified base quantity.  The actual quantity spawned is not the same as the entered quantity, but it is directly _proportional_ to it.

For disabling the retreat behavior, I guess an alternate could be added that gives them a galaxy-wide-chase-target of your home station or whatever (so they'll just beeline like an exo).

Toranth: yes, the exo costs currently go up exponentially with the power scale.  I want them to stop doing that (by being rolled into the relative strength thing).  Or, at least, not nearly as much of that.  Lancaster's Laws considered, something with 2x HP and 2x DPS should cost more than 2x the points.

Winge: on immunities, that shouldn't be factored in directly, if a particular bonus fleet type has an immunity and another doesn't, they're supposed to be balanced so that the one without the immunity still has roughly the same overall utility as the one with it.  Perhaps by having a different immunity, perhaps by being faster, perhaps by hitting harder.  Something like that.  If they're balanced, and they have the same cap, then their relative-strength values should be the same.

[Diazo]

I suppose I misread your last post from last night then.

Specifically, are the RelativeStrength values for the four golems valid or just place holders?

By valid I mean in use in game.

If they are "valid" or not in terms of that is what their relative strength should be is part of this thread.

D.

[keith.lamothe]

Specifically, are the RelativeStrength values for the four golems valid or just place holders?

By valid I mean in use in game.

They're used in some of the AI's "how much strength do I have on planet X" and "how much strength does the human have on planet Y" calculations.  In other such calculations, Firepower is used.  It's a bit of a mess, and if we can clean this one up that one's next.

So they're used, and I did intentionally set them, but I'm not sure that they're really a good indicator of how threatening a given golem is when compared to, say, starships or fleet ships.

If they are "valid" or not in terms of that is what their relative strength should be is part of this thread.

Don't take them as "these are right, don't change these".  For fleet ships and starships (again, anything the human can build normally, pretty much) I think those values are right and don't need to be changed in general (though some may need role-specific multipliers because they're more dangerous in exos, like zenith/spire ss's).  But the guardians and the golems and the spirecraft and the HKs and the mothership... those could be way off.  That's part of what I'm asking about here.