Knuckle Cracker

Particle Fleet => Gameplay Discussion => Topic started by: Bacteriophage on October 03, 2016, 04:27:42 PM

Title: Data on effectiveness of modules
Post by: Bacteriophage on October 03, 2016, 04:27:42 PM
I've done some testing, using custom maps and ships, on the effectiveness of various modules.  I may have miscounted a bit, but the numbers here should be mostly accurate.

Laser:
1 particle every 16 frames; 1.875 pps
1 energy every 30 frames; 1 eps
.533 energy per particle

Particle beam:
seems to target 1 particle per frame (could be faster)
1 energy every 10 frames; 3 eps
When trained on a single particle, the beam fires every 11 frames
.1 energy per particle slowed

Cannon:
1 particle every 40 frames; .75 pps
1 energy every 80 frames; .375 eps
.5 energy per particle

Missile:
1 particle every 61 frames; .492 pps
1 energy every 61 frames; .492 eps
1 energy per particle

MK7:
1 shot per 90 frames
1 energy every 5 frames; 6 eps
~12 particles per shot (hard to see, so this may be wrong)
1 particle every 7.5 frames; 4 pps
1.5 energy per particle

Fighter base:
2 fighters
1 particle per fighter per 8 frames; 7.5 pps
100 shots per fighter (give or take around 10 or so)
100 energy per fighter
7.5 energy per second
1 energy per particle

Grabber and Engine:
Neither seem to use any energy at all, and their effectiveness is not reduced by energy starvation.  Not what I would have expected, given the "depleted" look of engines on energy-less ships.

Lathe:
1 energy per second
1 struc built per second
1 energy per particle destroyed by struc
~500 frames (not precisely counted) to lathe a ship pickup or amp gem
~150 frames to lathe an energy pickup or energy mine
~667 frames to lathe an emitter
~2000 frames for an emitter to fully heal
These suggest that emitters have the same "health" as ship pickups, but heal themselves at 1/4 lathe rate.
~3333 frames to lathe a progenitor
Healing rate seems to be consistent with 1/4 lathe rate for emitters, suggesting 5X emitter "health" for progenitors.

Some shipbuilding observations:
When building a long "snake", each hull section builds every 61 frames and takes one energy (armored hull sections take 4x as long and use 4x the energy).  This is similar to (and probably exactly the same as) the rate when building multiple hull sections.  When building only one module (tested with laser and MK7), one energy every 15 frames is used.  This energy is not taken until the empty module part appears on the ship.  My (conjectural) model for shipbuilding is:
1. Build bridge at a rate of 1 energy per 15 frames, for the cost of the bridge in the editor.
2. Fill bridge with its 50 energy storage buffer.
3. Build hull out from center in steps of 61 frames, building all connected, unbuilt hull sections in each step at once.
4. During step 3, if a module outline appears, start a 2eps energy drain until the cost in the editor is paid.

Omnis:
Omnis carry 100 energy and can mire 285 land squares using that energy.  They mire once every 4 frames (note: the action of unmiring red mire and miring neutral land are equivalent for the following).  Thus we have:
.351 energy per particle
2.63 eps
7.5 pps
With the Omni cannon upgrade, omnis gain a cannon which shoots at a variable rate and energy consumption depending on efficiency:
1 shot per 30 frames -> 1 shot per 20 frames (2 pps -> 3 pps)
1 energy per 30 frames -> 1 energy per 40 frames (2 eps -> 1.5 eps)
1 energy per particle -> .5 energy per particle

Amp gem fire rate modification:
Laser: no effect on fire rate
Particle beam: no effect on fire rate
Cannon: 40 frames -> 28 frames (42.9% improvement)
Missile: 61 frames -> 43 frames (41.9%)
MK7: 90 frames -> 63 frames (42.9%)
Fighter base: 8 frames -> 6 frames (33.3%)
It appears that amp gems, when they have an effect, multiply fire delay by .7 and round to the nearest frame.  The rounding has no significant effect except for fighter bases, which are slightly less affected by amp gems.

Discharger:
100 frames to defeat a stunner
1 energy per 5 frames (6 eps)
20 energy per stunner destroyed

Hull damage:
Hull damage is difficult to test because a single particle may damage multiple hull sections, and I think this does affect energy usage (specifically, when testing particle damage against a square hull, the ship could stop only 86 particles before running out of energy, presumably due to multiple hull section hits.)  For damage of 1 square at a time, for both particles and struc, a ship, whether made of armor or normal hull, could stop about 100 particles before the ship ran out of energy (actually, the figures were 97 for particles and 101 for struc, but I'm going to ignore this discrepancy and use round numbers.)  Once depleted, an armored section took 8 hits and a normal hull section took 2 hits.  This, in combination with the ship rebuilding data, indicates the following stats for a constantly rebuilding hull section (either type):
1 energy every 61 frames (.492 eps)
2 particles per 61 frames (.984 pps)
.5 energy per particle

Shields:
The effect of shields is described here (http://knucklecracker.com/forums/index.php?topic=21680.msg149255#msg149255) by Virgil.  A shield uses 1 energy every 2 frames (15 eps).

Tactical implications:
First of all, we see that MK7s, particle beams, and fighter bases are the biggest energy hogs in the game, using significantly more energy per second than the relatively frugal cannon and missile.  The cannon shoots faster and takes less energy per second than the missile in addition to its greater versatility, leaving the only significant advantage of the missile being its homing capabilities.  Lasers never miss, are among the most energy-efficient weapons in the game, and are very powerful for their size, meaning that they can be quite effectively used offensively - as we have already seen with some custom ships.  Also, the MK7 is energy-inefficient for use on particles; it takes 3 times as much energy per particle as the cannon, while suffering an even greater chance to miss (with its longer range and all-at-once firing).  It may make a lot of sense to turn MK7's to "doppel and ship" mode when energy is a concern.  The omni's mire capability is extremely powerful (as powerful as a fighter base) and is the single most energy-efficient weapon in the game (the downside being, of course, the lack of flexibility of application of that power - only particles that actually hit land may be absorbed).  The omni cannon, while weak (1/4 the power) compared to the omni mirer, is essentially a double-fire-rate cannon module and thus can be decisive in particle fights.

Edit: added fighter base, corrected missile stats (thanks, Xalos!)
Edit 2: added engine, grabber, and lathe
Edit 3: added shipbuilding, omnis, and amp gem fire rate effects
Edit 4: added armored hull data, discharger
Edit 5: added hull damage, shields, and more tactical implications
Title: Re: Data on effectiveness of modules
Post by: Sorrontis on October 03, 2016, 04:34:34 PM
Beautiful analysis! I'm not sure if it's 100% right, but I agree that, simply by observation, the MK7 is not as efficient as it could be. That's why I don't use them on my ship designs (sometimes). I do use them against the players though.
Title: Re: Data on effectiveness of modules
Post by: Grabarz on October 03, 2016, 04:39:26 PM
Yup, I've certainly noticed the quirks with MK7 and particle beams. MK7 really is the ultimate ship killing weapon, but each shot disappears after one particle hit, meaning you're basically paying a ton of energy to destroy 3 particles. I've been switching my MK7's to ships only too. And yeah, I disable particle beams most of the time in the early game, they are a support weapon that really just wastes a ton of energy if you just let it fire when not needed.
Title: Re: Data on effectiveness of modules
Post by: Xalos on October 03, 2016, 04:48:56 PM
I've also done some tests, but for the lasers and particle beams I got wildly different numbers for energy usage (re-ran my experiments and got the same numbers as Bacteriophage):

Command Module: stores 50 energy, variable cost to build
Laser: costs 1 en/sec (0.533 en/particle, 1.875 particles/sec), range of 10 tiles, 10 energy to build
Cannon: cost 0.75 en/sec (0.5 en/shot, 1.5 shots/sec), range of 40 tiles, 15 energy to build
Missile Launcher: costs 0.492 en/sec (1 en/shot, 0.492 shots/sec), range of 45 tiles, 20 energy to build
Particle beam: costs 3 en/sec (0.1 en/shot, 30 shots/sec), range of 30 tiles, 15 energy to build
Energy Tank: stores 500 energy, 5 energy to build
Port: range of 50 tiles, 25 energy to build
Guppy: stores 100 energy, 25 energy to build
Reactor: generates 1.5 en/sec, 20 energy to build

I would have tested the energy consumption of cannons and missiles, but I'm stuck trying to figure out the CM cost equation (it's based purely on hull displacement, and it seems to be cubic, but I can't figure out the correct coefficients. :()

EDIT: Added stats for cannons, missile launchers, guppies, and reactors, as well as stats for shots/sec for weapon modules.  I also checked how long it takes a laser to kill a 1-health particle.  Missile launchers take 61 frames to fire rather than 60, and I'm not entirely sure why.

My analysis so far of lasers, cannons, and missile launchers:
Title: Re: Data on effectiveness of modules
Post by: Michionlion on October 03, 2016, 05:28:03 PM
So I recalled something about this way back in alpha, and at the time, this equation was being used (J was the creator, initially, I think): Cost=5+0.0005*(Hull-9)^2.5. That was over a year ago, however, so lots of things may have changed since then, and really Virgil is the final word on it.
Title: Re: Data on effectiveness of modules
Post by: Bacteriophage on October 03, 2016, 08:50:56 PM
Thanks for the data, Xalos.  I also counted 61 frames for the missile delay, but chalked it up to counting error.  I'm not sure what's going on with the particle beam and laser data, though.  I redid my experiment and got the same results.  One possibility: if you're frame-advancing on 4X speed, it skips 4 frames instead of 1...  Perhaps this is the cause of the discrepancy?
Title: Re: Data on effectiveness of modules
Post by: Xalos on October 04, 2016, 03:33:38 AM
Quote from: Bacteriophage on October 03, 2016, 08:50:56 PMOne possibility: if you're frame-advancing on 4X speed, it skips 4 frames instead of 1...  Perhaps this is the cause of the discrepancy?

I have no idea; I thought I did the tests at 1x speed but after re-testing I get the same results as you. It would explain both results being off by exactly a factor of four. In either case I've updated my stats with the correct numbers.
Title: Re: Data on effectiveness of modules
Post by: Tarmandan on October 04, 2016, 07:15:07 AM
I also did some testing just now (really should have seen if there was a thread up yet lol) but of the things I tested I came to similar conclusions.
Reactors generate 1.5 Eps
Tanks have a capacity of 500 E
Cannons take 0.5 E per shot
Command modules have an internal storage of 50 E

And I tested some AMP-gem stuff:
AMP-gems DO:
-increase range of all modules with a range (seems to vary, lathe got about 15-20% while discharger got around 40% increase, did not test further modules)
-increase fire rate of weapons (tested only cannons in the editor, but I think I noticed it on launchers and particle beams as well, so probably all weapons)
-reduce damage taken by armor (I did not test this, but virgil explained it in a post about shields)

AMP-Gems DON'T:
-increase lathing speed (tested with picking up AMP-gems)
-increase reactor output
-increase tank size
-increase built speed (and therefore repair speed)
-increase movement speed
-strenghten shileds( again, confirmed by virgil)

Not tested but might be affected: Energy consumption of modules, guppies, grabbers and fighters

Also, movement takes no energy but your ship still moves as if engine less if out of energy and ships can get a maximum of 30 EPS from a mine (again, not tested but stated by virgil)
Title: Re: Data on effectiveness of modules
Post by: Relli on October 04, 2016, 07:46:56 AM
It is SO nice to see this information start to crop up. Thank you to everyone who tested these out. I'd like to ask one question in particular that I've not seen anyone address yet. When building or rebuilding a ship, how is the energy used? What order do things get done in? Does each 1 energy go to a specific 1 thing? Is each 1 energy split into as many pieces as are needed at the time? Or is it perhaps that each 1 energy gives a full 1 energy to all systems at once? That'd be really awesome, but I can't imagine it's that one. Do ships build any faster if they have excess energy?
Title: Re: Data on effectiveness of modules
Post by: GoodMorning on October 04, 2016, 08:00:35 AM
Conjecture:
Energy goes into consolidated storage. Various systems then try to tap this. Hull builds in parallel, so it must tap as much power as possible up to some limit (per hull piece), and slow construction if insufficient energy is available. Modules likely work similarly.

So, I would guess that power is evenly split between all systems.

Note: It can't be one energy to all systems, otherwise a ship with multiple tank modules could farm energy without reactors.

As to build rate, more power only helps if there's a need for a lot at once after a lower-power period for the internal storage to charge. For example, bridge builds, 50 energy is requested over 50 frames (~1.66 sec), and hull starts. The eight or so tiles building when they become visible don't drain too much from storage (it's restored faster). As large ships expand, their power needs go up as more tiles are built simultaneously. At this point a full charge can give a small boost to construction.

So no, having 100 spare energy on board won't build any faster than having 6 spare energy. As long as there is enough power, build will go as fast as possible.



At least that's how I would have built it.

Edit: Possibly modules would consume in a queue, to fairly distribute nonfractional energy.
Title: Re: Data on effectiveness of modules
Post by: Tarmandan on October 04, 2016, 08:18:00 AM
For testing purposes I build a ship with only 3x9 tiles, and it had the command module in the back, in front of that a reactor and at the very front a tank. Reactor started building before the hull below the tank was even finished, tank still finished first. At least if enough energy is available all modules willl build at the same time and I assume the energy distribution is just as with the hull, equally spread out.
Title: Re: Data on effectiveness of modules
Post by: Relli on October 04, 2016, 09:54:52 AM
Quote from: GoodMorning on October 04, 2016, 08:00:35 AM
Note: It can't be one energy to all systems, otherwise a ship with multiple tank modules could farm energy without reactors.
I hadn't thought of that, that's a clever proof :D And yeah, I strongly suspected it didn't work that way. Otherwise gigantic ships would build MUCH faster than they currently do.
As for the rest of your post, thank you very much. It's nice to start actually understanding how this game works.
Title: Re: Data on effectiveness of modules
Post by: Bacteriophage on October 04, 2016, 10:19:26 AM
Quote from: Tarmandan on October 04, 2016, 07:15:07 AM
your ship still moves as if engine less if out of energy
Are you sure of this?  I set up a race between two identical ships with engines, one depleted of energy (from firing its missiles, which take exactly 1 energy each, so we know it is fully depleted), and the ships moved at the same rate.
Title: Re: Data on effectiveness of modules
Post by: Tarmandan on October 04, 2016, 10:25:48 AM
The engine thing wasn't something I tested but something I thought to have noticed during playing. Seems I saw wrong and probably got the engine killed the times I thought to "notice" this.
Title: Re: Data on effectiveness of modules
Post by: GoodMorning on October 04, 2016, 06:55:46 PM
For complete drain, hit it with a Stunner. Certainly that will drain all reserves.

Light but constant damage could also work to drain power completely, with the last hull repairs remaining incomplete as the race goes on.

It is not impossible that a ship will keep 1 energy in stock as a manoeuvering reserve, because a stopped ship is often a dead ship if it's already out of power range.
Title: Re: Data on effectiveness of modules
Post by: Xeneonic on October 04, 2016, 09:02:42 PM
Reading these values, we can constate that lasers deal, by far, the most damage out of any weapon when we take into account that 1 laser takes 1 cell worth of room on a ship. i.e. 4 lasers = 1 cannon/missile.

This is further amplified with lasers causing no overhead damage, cannot miss and is instantly transferred to the next target in range.
Title: Re: Data on effectiveness of modules
Post by: Relli on October 04, 2016, 09:28:45 PM
Quote from: Xeneonic on October 04, 2016, 09:02:42 PM
Reading these values, we can constate that lasers deal, by far, the most damage out of any weapon when we take into account that 1 laser takes 1 cell worth of room on a ship. i.e. 4 lasers = 1 cannon/missile.

This is further amplified with lasers causing no overhead damage, cannot miss and is instantly transferred to the next target in range.
I agree with all of this, save for the fact that I don't know what "overhead damage" means. Can you explain that for me please?
Title: Re: Data on effectiveness of modules
Post by: GoodMorning on October 04, 2016, 09:34:35 PM
Overhead = overkill. There is so much damage required to kill a particle, and any damage dealt over and above this amount is "overhead".

Ex: An MK7 to kill a single particle.

Better ex: Laser a particle until it is near-dead. Then hit it with a missile. Missile has done 1 damage to particle with 0.0001 health, making 0.9999 overkill damage.

Of course, sometimes it can be fun to kill a single Runner by hitting it with a battery of PZ Berthas. This has been called the "There is No Kill Like Overkill" mentality.
Title: Re: Data on effectiveness of modules
Post by: Relli on October 04, 2016, 09:52:57 PM
Thanks for the description. The example with the poor Runner cracked me up, so thank you as well for a good laugh :D
Title: Re: Data on effectiveness of modules
Post by: Bacteriophage on October 04, 2016, 10:39:30 PM
What really surprised me was the data on omnis.  Specifically, the omni cannon upgrade is both more effective and less effective than I thought.  More effective, because at max efficiency and omni acts as 2 cannons, meaning that when massed and upgraded, omni cannons are a significant addition to ships' firepower.  On the other hand, the base miring capability of omnis is 4 times as strong (as much as a fighter base), meaning that the upgrade doesn't make that much of a difference when taking mired land (also, this suggests that the base omni cannon target priority of always firing at mired land if there is some in range may not be the best use of resources - pointing the cannon at particles can be very effective, and the lack of the cannon won't cause the land to be mired that much slower).

Also, GoodMorning, I retested with a stunner, and the results were the same.
Title: Re: Data on effectiveness of modules
Post by: GoodMorning on October 04, 2016, 10:55:50 PM
Always good to check, and thanks for doing so.

Pointing the cannon at land is likely more effective, because no matter the health of the spacebourne particle, it can only mire one cell. So hitting a mired cell can to the effective damage of three shots for a hard particle, equal efficiency for a soft one, and wondrous efficiency for custom 5000-health particles.
Title: Re: Data on effectiveness of modules
Post by: Kharnellius on October 05, 2016, 12:01:55 AM
This thread needs a sticky.....badly.
Title: Re: Data on effectiveness of modules
Post by: GoodMorning on October 05, 2016, 12:04:16 AM
Once we have an API, we can PRPL-test it and put it on the wiki. So, by the time the data exists for certain, it will be placed elsewhere.
Title: Re: Data on effectiveness of modules
Post by: Tarmandan on October 05, 2016, 04:37:24 AM
Interesting, so Omnis have double the fire rate of a normal cannon at 100%? Then, has anybody got the data of how fast omnis mire land?

And it might be better to make a new post that has the top three posts reserved for data posting than to make this one sticky
Title: Re: Data on effectiveness of modules
Post by: planetfall on October 05, 2016, 06:58:08 AM
Quote from: GoodMorning on October 04, 2016, 09:34:35 PM
Overhead = overkill. There is so much damage required to kill a particle, and any damage dealt over and above this amount is "overhead".

Ex: An MK7 to kill a single particle.

Better ex: Laser a particle until it is near-dead. Then hit it with a missile. Missile has done 1 damage to particle with 0.0001 health, making 0.9999 overkill damage.

Of course, sometimes it can be fun to kill a single Runner by hitting it with a battery of PZ Berthas. This has been called the "There is No Kill Like Overkill" mentality.

It is my understanding that MK7 shots and normal cannon shots are the same thing. MK7 shots are easily observed as continuing through particles too weak to stop them. With regular cannons it's a bit harder but still visible. I may also have seen missiles do this as well, but it simply could have been a cannon shot going behind the missile.
Title: Re: Data on effectiveness of modules
Post by: PhailRaptor on October 05, 2016, 07:33:13 AM
Are we sure MK7 shots only kill one particle?  Because I routinely see the barrage of 3 bullets pass through multiple Particles before disappearing, usually 2 or 3 per bullet.  Although, I almost always have an AMP Gem on my Big Nose, so that could also play a role...
Title: Re: Data on effectiveness of modules
Post by: GoodMorning on October 05, 2016, 07:46:49 AM
MK7s can continue through multiple particles. I was taking an extreme example (extreme to the point that an MK7 will not target a single particle).

I understood that missiles, at least, were one-hit wonders. Still, the point remains: Overkill is the damage that is done over and above the requirement to kill the opponent.

Omnis can punch well above their weight in a high-particle health situation. Mire damages Doppels directly, making that different case, but can remove a very large number of "loose" particles, of any health, by first converting them to mire.
Title: Re: Data on effectiveness of modules
Post by: Bacteriophage on October 05, 2016, 08:05:12 AM
GoodMorning, I just tested, and unfortunately, hard particles mire 3 land cells, not 1.

Planetfall, I've tried firing cannons at 1-health particles and they don't pass through, but they destroy the particle.  But for particles damaged by lasers, they do pass through, suggesting exactly 1 soft particle of damage.  MK7 shots seem to do 4 or so damage (12 for the total barrage).

Tarmandan, I've put some more data into the first post, sorry if it's difficult to notice the change.  Anyway, omnis mire one land cell every 4 frames (7.5 pps).
Title: Re: Data on effectiveness of modules
Post by: GoodMorning on October 05, 2016, 08:41:17 AM
Hmm. Well, soft=1, med=2, hard=3+ health. So that suggests a different miring formula.

Do missiles keep going if not depleted?
Title: Re: Data on effectiveness of modules
Post by: Xalos on October 05, 2016, 12:59:24 PM
Quote from: GoodMorning on October 05, 2016, 08:41:17 AMDo missiles keep going if not depleted?

Missiles deal all their damage to a single target, so even if they hit a particle with 0.01 health left they will be destroyed.  Missiles that do not have a target any more will hang around where their target was for a while before despawning, and if a new target comes in range they will seek after it and attempt to destroy it.  In particular, this can be useful against emergent, because you might wind up with thirty missiles all hovering around the areas where new emergent are likely to spawn from, waiting to spawnkill any new emergent.
Title: Re: Data on effectiveness of modules
Post by: Sorrontis on October 05, 2016, 01:16:42 PM
I would love it if Virgil would shed some light on all this. I'm sure he has an algorithm somewhere he used to determine balance.  8)
Title: Re: Data on effectiveness of modules
Post by: powerofvoid on September 11, 2017, 10:19:20 PM
Since I couldn't find the answers anywhere: Engines

v ∝ E/n

v = ship speed going forward
E = number of engines
n = number of hull tiles placed


"n" is the number used to calculate CM cost, which is not affected by:
Unfortunately, there is no way to see "n" directly in the ship editor -- you would need to either count directly in the editor, or reverse-engineer the CM cost formula.

EDIT: Unless your hull consists entirely of unarmored square tiles.