Logo Platform
logo amplifiers simplified

To Glass Cannon or Not to Glass Cannon

Reply
Copied to clipboard!
11 years ago
Jul 29, 2013, 1:24:20 AM
Or we could try to come up with a complete description of all the mechanics and read out all descriptors for hulls/cards/weapons/defenses/tonnage.

But this would end up in a huge undertaking, either way. The autobattle is, more or less, what we'd need, with an input scheme for all the ships that enter the fight. It should exist in the game engine. Certainly, it won't be a copy-paste-job, as you'd need some descriptor input or even GUI for actually running the simulator, but the main combat code has to be there, already.



Since it's easy enough to come up with the expected damage for any kind of weapon loadout versus some certain defense, the only problem here is complexity due to the number of combinations and variations in total. So what I'm thinking about is actually running, let's say, a 100 fights in sim per combination of weapon/def/hull as a target, leaving cards aside. You'd end up with a three-dimensional plot for each type you're testing versus everything else, entering a point with a weight of the percentual damage the ship has inflicted to the other ship. You could of course do a second plot of this kind for damage taken.



Anyways, this would already be a quite reduced problem and would still end up forcing you to define several reference types of ships: Omnidefense, specialised defense each type, long/middle/short range of each weapontype, for all tech variations and hulltypes that are viable.



In the end, you'd then need to define some nonarbitrary state for the game being balanced. I'd put forward, that each archetype you come up with needs to lose in some areas of the phasespace you created and to win in others. If there's any of the 3D-plots where you end up "winning" in >75% of all cases or "losing" in >75% of all cases, I'd call the game unbalanced.



So this is the scale of randomized testing I'd like there to be in the future for balancing the game, to ensure quality. But maybe I'm biased towards using massive processor power statistical analysis instead of rational thought, because I disbelieve this can be solved analytically by anyone but a genius to everyone's content. smiley: wink



thuvian wrote:
We could jurry-rig a combat simulator, I've got something like that setup for myself now. (It's a save game of a Harmony game with infinite production in a big galaxy with my personal WIP mod.)

To get something fairly function we'd need:

  • A custom galaxy with 1 star per Empire all connected to a single center star. (I don't know how)
  • A mod to the Empires to make them produce infinite (or just a bunch of) FIDS each turn. (I have this already)
  • A mod to the Empires to remove strategic resource requirements. (I know how to do this)
  • A mod to turn off the AI building and fleet movements ( I really don't know how to do this)





To "run" the simulator.

[list=1]
  • Start or load a game with this mod.
  • Pick the first Empire that you want to run a simulation with.
  • Design the Ships for that Empire and set the empire to build them.
  • Also design and build some random trash ships.
  • End Turn.
  • Create 1 fleet with your desired skirmish side.
  • Create several other fleets with the random trash fleets.
  • Send the fleet of interest to the center.
  • Save & Quit.
  • Edit your save file to remove your SteamID.
  • Restart ES and load your game.
  • Choose the Empire for the other side of the combat.
  • Design the Ships for that Empire and set the empire to build them.
  • Also design and build some random trash ships.
  • End Turn.
  • Create 1 fleet with your desired skirmish side.
  • Create several other fleets with the random trash fleets.
  • Send the fleet of interest to the center.
  • Send the random trash fleets to the opponent Empire's homeworld.
  • Save the game.
  • Run the combat of interest.
  • Load game.
  • Run 1 trash fleet combat to reset our place in the RandomSeed.
  • Run the combat of interest.
  • Repeat as necessary.

  • [/list]



    Once the mod is setup, it wouldn't take much to actually run the tests. The majority of the time would be starting up ES so many times.
    0Send private message
    11 years ago
    Jul 29, 2013, 3:43:17 AM
    Just use some C++ or your favorite code of choice, trying to mod the game into a combat simulator is an exercise in futility. At least as of v1.1.14
    0Send private message
    11 years ago
    Jul 28, 2013, 11:47:13 AM
    There has been much discussion on the viability of ships other than glass cannons. An exhaustive comparison of all Empires, hulls, technologies, heroes, and upgrades is beyond any single post. However, we can simplify the situation and examine a small set of cases and some general principles.



    In this post I take the stats from 1.1.14 as a basis of the discussion.



    First, we need to have an overview of the various modules that will play a roll.



    [CODE]Armor Technology

    level weight cost MaxHealth Defense

    1 .15 16 0.5 .05

    2 .15 24 0.75 .1

    3 .15 36 1.00 .15

    4 .15 60 1.25 .2

    5 .15 80 1.50 .25

    6 .15 110 1.75 .3

    Increases by 0.25% of max health per level (additive)

    Increases by 0.05% of Defense per level (additive)[/CODE]





    [CODE]Defense Technology

    Defense Cost Wt IA D A D

    Kinetic 1 4 8 0 3 0 50

    Kinetic 2 16 8 0 5 0 60

    Kinetic 3 70 8 0 10 0 120

    Shield 1 4 6 0 0 10 75

    Shield 2 24 6 0 0 50 100

    Shield 3 70 6 0 0 150 150

    Flakk 1 8 4 10 0 0 50

    Flakk 2 24 4 17 0 0 60

    Flakk 3 70 4 34 0 0 90

    [/CODE]



    [CODE]Weapon Technology

    WEAPONS tdam netdam netdam.wt

    4 Kinetic1LR 350 525.00 58.33333

    12 Kinetic2LR 1050 1575.00 175.00000

    17 Kinetic3LR 1950 2925.00 325.00000

    6 Kinetic1MR 560 929.60 84.50909

    14 Kinetic2MR 1600 2656.00 241.45455

    22 Kinetic3MR 3600 5976.00 543.27273

    18 Kinetic1SR 1152 2304.00 329.14286

    13 Kinetic2SR 680 1360.00 194.28571

    21 Kinetic3SR 1840 3680.00 525.71429

    1 Laser1LR 210 315.00 28.63636

    5 Laser2LR 450 675.00 61.36364

    8 Laser3LR 900 1350.00 122.72727

    11 Laser1MR 600 996.00 142.28571

    19 Laser2MR 1680 2788.80 398.40000

    26 Laser3MR 4512 7489.92 1069.98857

    3 Laser1SR 240 480.00 53.33333

    9 Laser2SR 576 1152.00 128.00000

    23 Laser3SR 2880 5760.00 640.00000

    7 Missile1LR 440 660.00 94.28571

    16 Missile2LR 1320 1980.00 282.85714

    25 Missile3LR 3960 5940.00 848.57143

    10 Missile1MR 720 1195.20 132.80000

    20 Missile2MR 2640 4382.40 486.93333

    27 Missile3MR 8200 13612.00 1512.44444

    2 Missile1SR 224 448.00 40.72727

    15 Missile2SR 1360 2720.00 247.27273

    24 Missile3SR 4000 8000.00 727.27273

    [/CODE]

    WEAPONS - Weapon Name

    tdam - Single Phase Damage (ignoring accuracy)

    netdam - Total Damage given out over all Range Phases

    netdam.wt - Total Damage across all Range Phases divided by Weapon Weight



    Notice, Medium Missile Level 3 are insanely good. I believe it is an error in the dev's formula, bit it is accurate according to the game files.



    From this we can develop the load out of the basic Glass Cannon Missile Destroyer.

    [CODE]

    LRM 1 Destroyer 6160 alpha 9240 damage 450 health $82

    LRM 2 Destroyer 18480 alpha 27720 damage 450 health $166

    LRM 3 Destroyer 55440 alpha 83160 damage 450 health $418

    [/CODE]



    Alpha is the damage dealt only in the Long Range Phase. For a counter ship to survive this is the damage it must be able to soak.

    damage is the total damage dealt across all 3 Range Phases



    Now, I'm going to ignore accuracy, deflection and the rest per the other threads about them, they really don't matter in large scale engagements and only make a difference in small ship engagements with few weapons. These ships are packing 14 Long Range Missiles with 100% accuracy, they aren't going to be suffering in damage.



    So, to create a ship to counter the LRM1 we need to be able to soak 6160 damage and be mindful of the $82 price tag.



    We can create a combination of all possible combinations of level 1 technologies with destroyers to see what options we have.

    I'm going to use the Harmony Destroyer as our basis.



    [CODE]

    ShipID armor def wt mdef health hw ehp weapons dam.alpha dam.net cost

    41 5 5 95 312.5 1575 100 6496.875 0 0 0 160

    121 1 17 83 892.5 675 100 6699.375 2 880 1320 198

    87 2 12 78 660.0 900 100 6840.000 3 1320 1980 177

    67 3 9 81 517.5 1125 100 6946.875 2 880 1320 166

    54 4 7 88 420.0 1350 100 7020.000 1 440 660 163

    128 1 18 87 945.0 675 100 7053.750 1 440 660 203

    94 2 13 82 715.0 900 100 7335.000 2 880 1320 182

    135 1 19 91 997.5 675 100 7408.125 1 440 660 211

    48 5 6 99 375.0 1575 100 7481.250 0 0 0 168

    74 3 10 85 575.0 1125 100 7593.750 2 880 1320 174

    142 1 20 95 1050.0 675 100 7762.500 0 0 0 216

    61 4 8 92 480.0 1350 100 7830.000 1 440 660 171

    101 2 14 86 770.0 900 100 7830.000 2 880 1320 190

    149 1 21 99 1102.5 675 100 8116.875 0 0 0 224

    81 3 11 89 632.5 1125 100 8240.625 1 440 660 179

    108 2 15 90 825.0 900 100 8325.000 1 440 660 195

    68 4 9 96 540.0 1350 100 8640.000 0 0 0 176

    115 2 16 94 880.0 900 100 8820.000 0 0 0 200

    88 3 12 93 690.0 1125 100 8887.500 1 440 660 187

    122 2 17 98 935.0 900 100 9315.000 0 0 0 208

    75 4 10 100 600.0 1350 100 9450.000 0 0 0 184

    95 3 13 97 747.5 1125 100 9534.375 0 0 0 192[/CODE]



    [CODE]

    Table Key

    ShipID Unique Ship number to keep track of them

    armor # of armor modules on the ship

    def # of defense modules

    wt # total weight of the ship before the weapons are added to fill the leftovers

    mdef # total missile defense (including armor module bonus)

    health base health, 450 + armor bonus

    hw HullWeakness

    ehp Effective Hit Points

    weapons # of weapon modules

    dam.alpha Damage in the Long Range Phase

    dam.net Damage across all Range Phases

    cost How much the ship costs to build[/CODE]



    Here are the 21 highest Effective Hit Point destroyer builds. At this point, you might notice that lots of them have more than 450 dam.alpha and more than 6160 Effective HP. Great, so that means Glass Cannons are weak, right?



    Well, no. I set you up to fail there. Sorry.

    1. Hull Weakness for destroyers in 1.1.14 is actually 300, so that means the effectiveness of defenses is only 1/3 of what is listed here. I think this is an oversight by the devs, so I'm going to ignore this one, and let you have an HW of 100 for free.

    2. This is balanced on a per ship basis, notice that the price tag for all of these is much higher. If our defender ship is attacked by 2 Glass Cannon LRMs, it dies, but only takes 1 Glass Cannon with it, so the Glass Cannons win.

    3. It's actually even worse. This only takes into account 1 defense. What if you aren't up against 1 type of Long Range Glass Cannon, but a mix of the three different long range weapons. Our EHP is then only 450 against the other two types. To counter those we'd need to have equal amounts of all 3 defenses. Notice that we can barely defend against 1 weapon type, against 3 there is no possible way to defend in a cost effective manner.



    What about Dreadnaughts you say? Well, I did the same thing using the Harmony Athenic (I didn't use the -75% weapon weight to make things fair for other Empires. However, more weapons would just make the ship more dangerous not safer. Also, you'd want Medium or Short Range weapons to counter the Glass Cannons, but again we'll ignore that.)

    [CODE]

    ShipID armor def wt mdef health hw ehp weapons dam.alpha dam.net cost

    430 2 61 364 3355.0 6000 100 207300.0 5 2200 3300 1335

    264 4 37 388 2220.0 9000 100 208800.0 1 440 660 1163

    333 3 47 368 2702.5 7500 100 210187.5 4 1760 2640 1236

    437 2 62 368 3410.0 6000 100 210600.0 4 1760 2640 1340

    444 2 63 372 3465.0 6000 100 213900.0 4 1760 2640 1348

    271 4 38 392 2280.0 9000 100 214200.0 1 440 660 1171

    340 3 48 372 2760.0 7500 100 214500.0 4 1760 2640 1244

    451 2 64 376 3520.0 6000 100 217200.0 3 1320 1980 1353

    347 3 49 376 2817.5 7500 100 218812.5 3 1320 1980 1249

    278 4 39 396 2340.0 9000 100 219600.0 0 0 0 1176

    458 2 65 380 3575.0 6000 100 220500.0 2 880 1320 1358

    354 3 50 380 2875.0 7500 100 223125.0 2 880 1320 1254

    465 2 66 384 3630.0 6000 100 223800.0 2 880 1320 1366

    285 4 40 400 2400.0 9000 100 225000.0 0 0 0 1184

    472 2 67 388 3685.0 6000 100 227100.0 1 440 660 1371

    361 3 51 384 2932.5 7500 100 227437.5 2 880 1320 1262

    479 2 68 392 3740.0 6000 100 230400.0 1 440 660 1379

    368 3 52 388 2990.0 7500 100 231750.0 1 440 660 1267

    486 2 69 396 3795.0 6000 100 233700.0 0 0 0 1384

    375 3 53 392 3047.5 7500 100 236062.5 1 440 660 1275

    493 2 70 400 3850.0 6000 100 237000.0 0 0 0 1392

    382 3 54 396 3105.0 7500 100 240375.0 0 0 0 1280

    389 3 55 400 3162.5 7500 100 244687.5 0 0 0 1288[/CODE]



    Now the average cost for any of these is around $1200. How many $82 destroyers is that, a little over 14!. What's the alpha strike damage for 14 Glass Cannon LRM1 Destroyers? 86,240 damage. That's a big chunk. How many destroyers can our dreadnaught kill per combat? With long range weapons, we'd only kill 1 per Range Phase, or 3 total ships. That's not going to get us very far. With Short Range weapons we could kill 12, but we'd need to do an alpha strike of 3600, but that damage HAS to hit. Using short range weapons (which we have to, in order to kill them fast enough) and with the horrible accuracy of short ranged weapons (.5 for kinetics) is balanced by the extra damage of kinetics (300% in some sense compared to LRMs, I think this was a typo too, but it is in the files).



    Now, the numbers are still looking pretty good, but again we have to remember that this is only for a single defense, if we split the defenses across all three defense and use ship 389, the beefiest one, our EHP drops to 86,605. Oh... Remember how much damage our alpha strike Glass Cannon Fleet was doing? It was 86,240. That's a countering.



    Worse, is that our Glass Cannon destroyers could just add 1 defense of each type, their damage would drop by 15%, price would increase to $100, but we'd need to do double damage against them to kill them. Which we wouldn't be able to do in time. They'd end up killing our dreadnaught in the Medium Range Phase (doing 76k damage), and we might be able to get 4 of them, giving the Glass Cannon Destroyers $800 on the battle.



    In summary, I'm pretty sure Glass Cannons are the king.



    Caveats.

    I did some simplifications to make this manageable. The efficiency of the weapons obviously differs. When you have a technology level advantage, there will be obvious differences. At higher levels of technologies you might think the outcome will be different, especially given how armor adds a % increase and is cumulative & healing modules. However, weapons become 900% more powerful, but defenses only increase by 200%!!!. The prices for defenses increase by 1000% to 2000% whereas the prices for weapons increase by only 900%. So weapons are getting more powerful more cheaply compared to defenses. Bombers and Fighters would require the Dreadnaught to invest into a 4th defense type (I'm not sure how deeply). Prices are listed in industry, but many races would be able to convert dust to industry at a better rate drastically change the price relationships. Heroes and Empire Bonuses can drastically change the math involved, but that goes both ways. The Harmony actually has the best end game glass cannons due to Metal Memory giving them hero bonuses on every single ship, and not just their main fleets. Plus, they can't "loose" their heroes as easily.



    The counter

    The counter to Glass Cannons are cheaper Glass Cannons, spend just $60 on a ship with a few weapons. Both will die, but you spent less.

    The counter to that is ships with a few more defenses. They kill your cheap glass cannons and still survive.

    This countering continues until a sort of equilibrium is reached, either in weapon:defense rates on all ships, or in the composition of the various fleets.

    Or you just don't worry about it and continue to just throw cheap ships at each other in a game of economic chicken.



    So, what does it all mean?

    It means that we need to spend less time obsessed with our own personal experiences in the game, especially against the AI. We need to consider playing evidence along with math that demonstrates and matches our experiences. To rely on either on exclusively when balancing leads to false conclusions.



    Finally, have some pictures of what happens when you remove the special module requirement from bombers.

    0Send private message
    11 years ago
    Jul 29, 2013, 7:50:36 AM
    In my opinion simulation comes second and is just a tool to verify that it works like you expect it to work.

    The very first step really should be making up your mind on what exactly you expect!



    And seeing that we can freely adjust these values via modding, we could actually use the devs holiday-week to make a "proposal".



    As in: We first define what we expect to happen, then make a mod that should realize it (if our understanding of the combat-mechanics is close enough to how it works) and at the end send it to Meedoc and ask about his thoughts on it. If we can explain our thought-process and have several people agree on it, I think the chances are good that our changes could be adapted.
    0Send private message
    11 years ago
    Jul 29, 2013, 8:02:07 AM
    That sounds good to me.



    My $0.02.



    Problem: Combat isn't very interesting or deep.



    Background:

    ES takes the revolutionary step of mixing strategic combat with just a touch of tactical combat. Unlike current Turn-based + Real Time Combat titles, the control over combat in ES is limited to general commands, emulating the sort of strategy a fleet admiral might deploy in a real fleet battle. These commands take the forum of cards or tactics that can be deployed during battle. Combat outcomes are then dedictated primarily by fleet composition and ship design and less so by the cards.



    Card Battle

    During battle there is typically a few cards to play that is good for you, some that are okay for you, one that counters what would be really bad for you. A conservative card strategy would be to play a card that neutralizes your opponent's card, and let the outcome of the battle depend purely upon the fleets involved. Such a conservative approach is sensible, but isn't interesting or fun. It takes a rock-paper-scissors system and boils it down to always choosing Rock because because paper can't hurt you (much). While I believe there is a large potential here for interesting outcomes, this topic will not be the subject of this thread.



    Fleet Composition, Ship Design, and Combat Mechanics.

    These are the fundamentals of the combat system and each one is affected by the others that fall before it. To make the game interesting we should focus mostly and combat mechanics and see how those changes filter through the system.



    Current Combat Mechanic Overview

    Accuray - Base chance to hit

    Evasion - Chance to evade

    EvadeDisorientation - How much evasion is lost per succesful evade, currently 0.15

    Evades - How many succesful evades were made in this round of combat by this ship

    To Hit Formula - Accuracy*AccuracyModiers - (Evasion*EvasionModifiers - 0.15*evades)

    Combat - consists of 3 Range Phases

    Range Phases - in 1.1.14 accuracy is modified by your Weapon Range, in 1.1.9 it was a gross damage modifier

    Weapon Range - Melee (i.e., Short), Medium, Long

    RoundsToReload - How many additional rounds are required to prepare a weapon to launch another salvo, long = 3, medium = 2, short = 0

    RoundsToHit - How long it takes a weapon to hit, long = 3, medium = 2, short = 0

    Weapon Fire Rate - Currently set per type, Short = 4 times per round, Medium 2 times per round, Long = 1 time per round

    Weapon Type - Kinetic, Missile, or Laser. Each is affected by only its matching defense type

    MinDamage - The minimum damage dealt by the weapon upon a succesful hit

    MaxDamage - The maximum damage dealt by the weapon upon a succesful hit. Currently MinDamage=MaxDamage

    CritDamage - The extra damage dealt due to a critical hit

    CritChance - The chance of dealing a critical hit

    ShotsPerSalvo - The number of attacks per salvo. They either all hit or all miss.

    Damage = ShotsPerSalvo * MaxDamage

    Defense Type - Deflectors (vs Kinetic), Shield (vs Laser), and Flakk (vs Missile)

    Hull Weakness - To increase Effect Hit Points 100%, this is the required Defense Strength

    Hit Points - The life of the ship, when this reaches 0, the ship is dead

    Effect Hit Points (EHP) - A transformation to Hitpoints based upon armor & defenses that allow easier comparisons of the 3 damage types

    Hull Size - Each Hull has an associated cost, weight, special module limit, Command Point Cost, Hit Points, and Hull Weakness.

    Weight - Each hull has a specific weight which is the weight amount of modules that can be added

    Special Module Limit - Troops, Fighters, Bombers each use up 1 or more special module slots. 4 is the maximum special modules per hull

    Cost - The cost to produce a ship is measured in industry points. Upgrading via dust is also possible for most Races.

    Repair - Some modules grant a repair value either a flat amount of based on the Maximum Hit Points of a ship

    Command Points - Each Hull has a Command Point Cost, larger ships require more Command Points. Fleet size is limited by Command Points

    Fleet - A group of ships that fight together during battle

    DeflectionsPerRound - Number of kinetic salvos that are deflected each round without damage.

    Absorption - Amount of Laser damage blocked each round

    InterceptionAccuracy - Some form of missile defense. Classic ES notes suggest that each Flakk fires 1 each round and has an X chance to destroy a missile. No idea how it is implimented in ESsmiley: biggrinisharmony

    Fighter - A special module unit, a group of space fighters accompany this ship and will engage fighters, bombers or ships. Unopposed it will make 0/2/2 attacks during the Long/Medium/Short Range Phases. In enemy fighters or bombers are present then it does something else.

    Bomber - A special module unit, a space bomber accompanies this ship and will only engage enemy ships. It will make 0/2/2 attacks during the Long/Medium/Short Range Phases.

    Fighter Defense - A defense against Fighter and Bomber craft. It has a special module. It also is increased for the amount of other defense modules on a ship.

    Ship Bonuses - These bonuses affect only the ship it is on. These stack with themselves and other bonuses.

    Fleet Bonuses - These bonuses affect all ships in the Fleet. These stack with themselves and other bonuses.



    Current Combat Aspects

    Accuracy, Evasion, Deflection, Interception, and Absorption are typically minimized with a large number of attacks. They are very shallow mechanics.

    Range Phase - Currently it affects the accuracy of weapons based upon Weapon Range. Since Accuracy is already minimal it also has a minimal impact.

    Hull Weakness - Currently little ships have a hull weakness of 300 which results in defenses being 33% effective. It really makes defense pointless on little ships.

    Passive Modifiers on Defense Modules. These are stackable and result in largely unbalanced effects.

    Long Range Missile Destroyers offer the best offense for the least cost. Larger ships are rendered mostly obsolete by them due to an inefficiency in allocating weapon damage efficiently, the ability for missile weapons to always launch their most effect salvo, the rapid scaling of weapons, the necessity of maintaining all three defenses, among other things.

    All three defenses are required on a ship, because if you are missing one category, your EHP is effectively your base HP, which means that a single weapon may be sufficient to destroy your ship.

    Long Range Missiles are currently the slowest weapon, so that the missiles may be attacked by the missile defenses 4 times. In Classic ES, 4 x number of flakk modules worth of missile could be destroyed in a round. This is less effective in Disharmony although the exact mechanics are unknown. However, Medium & Short Range missiles "travel" much faster.

    Weapon Speeds. In order to balance weapons, speed was added, which makes no sense. Speed does not depend upon Weapon Type, but instead, range type! Short range weapons are faster?! than long range weapons. This means that you can counter Long Range Destroyers with short range weapons, a solution that is very dissatisfying.

    Combat Ranges - In order to add depth to combat, range was added, however it doesn't make a lot of sense. Instead of the distance between the ships increasing, weapon range is implimented by changing how quickly shots each your opponents. Short range shots travel faster than long range shots. It is understandable from a balance perspective, but very dipleasing from a aesthetic perspective.



    Fighter and Bomber Mechanics

    I don't know very much about these, so I will not comment. Any balance changes that effect this portion of combat are unintended.



    Current Proposal

    Drop all current passives from Defense Modules except for Fighter Defense modifiers

    Change Hull Weakness back to 100 for all ships

    Change accuracy for weapons, Lasers +25%, Missiles + 50%, Kinetics + 0%.

    Change damage for weapons, Lasers +25%, Missiles + 0%, Kinetics + 50%.

    This changes overall efficiency for the weapons to Lasers +56%, Missiles +50%, Kinetics +50%.

    Set EvadeDisorientation to 0. This makes accuracy a valid defense type. Small Hulls have accuracy 0.50, medium ?, large 0.15.

    Weapon Speed should be Lasers, Missiles, Kinetics (in descending order).

    Reduce ShotsperSalvo to more reasonable numbers, preferably less than 10 (or 20). Adjust damage as necessary.

    Change Range Phase modifiers to accuracy. For Long Range -50%, Medium Range - 25%, Short Range -0%. I don't know if this should be additive or multiplicative.



    Combat Formations

    Currently, combat is modelled after ye olde early modern combat line formation. The combat simulation makes it look like a line of battle formation, but that's misdirection. In a line formation, each ship is arranged in line "shoulder to shoulder" as it were, facing the enemy in a similar formation. Instructions are then given to either NoseBreaker ("everyone shoot the guy on the far left") to Guillotine ("everyone shoot the three guys on the far left") to Spreadfire ("shoot the guy in front of you") (once all targets have someone shooting at them, then the process is reapplied, so that all target have roughly equal numbers of attackers). In defense, your choice is how gets to line up on the left first? Either by defense total, hit point total, weight total, or least siege.



    Obviously, per the Early Modern tactics (e.g., Napolean Warfare, American Revolution) troops fought using a line (and columns, and square too) formation, it's not very interesting, very bloody, and has very predictable results. It is strong in that it allows maximimum firing width and concentration and movement (at least in certain directions). Howver, it also is weak against flanking, rear attacks, and guys hiding in bushes and not engaging in fair fights. However, none of these options are available in ES, everyone fights "like a civilized being". Everyone lines up and then slowly walks toward each other while shooting. Changes in this, I believe could be the source of a lot of interesting mechanics. Things like pincer formations, envelopment, flanking, etc could prove very interesting.



    Naval tactics through the ages would also provide useful analogies upon which to base game mechanics.



    We can look at Battleship era (WW1, early WW2 even) to see the culmination of Big Ship tactics. I'm not very versed in this, but it seems that the bigger ships were used to hold the bigger guns. Bigger guns meant longer range. Longer range meant you could start shooting and damaging your opponent first. The second concern was speed (the BattleCruiser doctrines combined speed & firepower). To model this sort of design we would need to:

    1. Develop Weapon Technologies with truly obscene weight requirements. These weight requirements would forbid them being used on smaller ships.

    2. Develop more Range Phases. If we have a Super Battleship with the Yamamto Cannon I was to start firing from Absurdly Far Range, not just in Long Range like the rest of the peasants.



    However, this was done away with with the advant of Air Power and the ability to project power even more cheapily from an even farther distance. In this case, the ideal battle plan is drop off many spacefighters and hope some of them can get to the enemy carrier. Your carrier you don't risk in battle. Modelling the game after Carrier Based Tactics could also prove interesting.



    Another option is to start our Tactics plans off with early ship tactics and end up developing Carrier Based Tactics. Then progression through the technology tree would reflect changes in both ship design, fleet composition, and battle strategy. I think this has the potential to be the coolest, but it also would require the most thought and work.



    Part of the problem in specifying a good combat system is the lack of focus on what style of combat we are reflecting. If we want BattleShip Doctrines to rule the day, then we acknowledge that and try to make sure the mechanics fit it.
    0Send private message
    11 years ago
    Jul 29, 2013, 9:14:35 AM
    So instead of continuing to just talk, I'd like to propose my suggestion here.



    First let me define how I imagine an average ship:



    My average ship wears 1 armor-module of the best type (15% tonnage), uses 35% tonnage on defenses and the remaining 50% tonnage on weapons.



    1. I would expect it to barely destroy a ship similar to itself that has no defenses against the used weapon-type.

    2. I would expect it to do roundabout 50% damage to a ship similar to itself that has omni-defense.

    3. I would expect it to do roundabout 25% damage to a ship similar to itself that specialized it's defenses against it.



    For the sake of progression 1. does not need to be true in the lategame anymore but 2. and 3. still should.



    So some examples:



    An early-game 1 CP ship with 1 Armor-Module has 787.5 HP and 100 tonnage.

    In order to barely destroy it with 50% on weapons the damage per tonnage needs to be 16.

    Assuming a Hull-Weakness of 100, in order to half the damage taken by 12 tonnage the defense needs to be roundabout 9 per tonnage.



    A mid-game 1 CP ship with 1 Armor-Module has 1012.5 HP and 162.5 tonnage.

    In order to depict progression in both weapons and defenses, I'd allow a ship to be strong enough to kill 2 of these with 50% weapons.

    So let's say 2050 damage which comes to 25 damage per tonnage.

    Now we need 19 tonnage to offer enough defense to cut initial damage down to 1/4th. That's 300 defense needed and thus 16 defense per tonnage.



    A late-game 1 CP ship with 1 Armor-Module should come out at 1350 HP and 250 tonnage.

    This time I'd say the weapons should be able to destroy 3 of these if undefended. So 4200 damage. That's 34 damage/tonnage.

    Now we have 30 tonnage to reduce the initial damage to 1/6th that's 20 defense per tonnage.



    Does this also work for Dreadnoughts?

    Let's see:

    5400 HP, 900 tonnage.

    Damage: 450x34=15,300 damage against no defenses. So here it is okay, as expected.

    However, we would have 105 tonnage for armor and here we'd be able to get a whoppin' 1680 defense, which means damage is cut down to 1/17th instead of 1/6th.



    So I'd argue, that defense-modules should probably consume a percentage of the hull, as their benefit, like with armor, is adding a percentage in effective health.

    I'd say 6%-steps for 50/150/250 defense for the different levels would be in line with the damage/tonnage as shown above.



    When it comes to differnces between ranges.

    Here's my suggestion:

    LR=75% in Phase 1, 20% in Phase 2, 5% in Phase 3 for a total of 100% damage

    MR=40% in Phase 1, 40% in Phase 2, 40% in Phase 3 for a total of 120% damage

    M=10% in Phase 1, 40% in Phase 2, 100% in Phase 3 for a total of 150% damage



    So Long-Range is for dishing out most damage initially and then having to deal with less opponents later.

    Medium-Range is for a constant-damage-output over all phases.

    Melee does most damage but needs to survive the first turn.





    For the difference between missiles/beams and kinetics I would go back to what it was in Vanilla but not with the ranges.



    I'd say:

    Missiles: Only hit once per battle-phase but do 120% of the damage.

    Beams: Hit twice per battle-phase for 110% of the damage.

    Kinetics: Hit 4 times per battle-phase for 100% of the damage.



    This way the choice would not be that big of a deal but depend on what ship-sizes you are fighting. Aquiring up to 4 targets per turn is an advantage when fighting mostly small ships, so you won't overdamage them so badly. If fighting less targets going for more damage seems to be the best idea.
    0Send private message
    11 years ago
    Jul 29, 2013, 10:51:30 AM
    [CODE]

    ShipID armor def wt mdef health hw ehp weapons dam.alpha dam.net rounds2die rounds2kill rounds2killwshort rounds2killwmissile rounds2killwshortmissile

    1 0 0 0 0 450.0 100 450.0 14 1176 1568 0.5739796 1.004464 0.6696429 0.8370536 0.5580357

    2 1 0 15 0 787.5 100 787.5 12 1008 1344 1.0044643 1.171875 0.7812500 0.9765625 0.6510417

    8 0 1 35 100 450.0 100 900.0 9 756 1008 1.1479592 1.562500 1.0416667 1.3020833 0.8680556

    3 2 0 30 0 1125.0 100 1125.0 10 840 1120 1.4349490 1.406250 0.9375000 1.1718750 0.7812500

    15 0 2 70 200 450.0 100 1350.0 4 336 448 1.7219388 3.515625 2.3437500 2.9296875 1.9531250

    4 3 0 45 0 1462.5 100 1462.5 7 588 784 1.8654337 2.008929 1.3392857 1.6741071 1.1160714

    9 1 1 50 100 787.5 100 1575.0 7 588 784 2.0089286 2.008929 1.3392857 1.6741071 1.1160714

    5 4 0 60 0 1800.0 100 1800.0 5 420 560 2.2959184 2.812500 1.8750000 2.3437500 1.5625000

    6 5 0 75 0 2137.5 100 2137.5 3 252 336 2.7264031 4.687500 3.1250000 3.9062500 2.6041667

    10 2 1 65 100 1125.0 100 2250.0 5 420 560 2.8698980 2.812500 1.8750000 2.3437500 1.5625000

    16 1 2 85 200 787.5 100 2362.5 2 168 224 3.0133929 7.031250 4.6875000 5.8593750 3.9062500

    7 6 0 90 0 2475.0 100 2475.0 1 84 112 3.1568878 14.062500 9.3750000 11.7187500 7.8125000

    11 3 1 80 100 1462.5 100 2925.0 2 168 224 3.7308673 7.031250 4.6875000 5.8593750 3.9062500

    17 2 2 100 200 1125.0 100 3375.0 0 0 0 4.3048469 Inf Inf Inf Inf

    12 4 1 95 100 1800.0 100 3600.0 0 0 0 4.5918367 Inf Inf Inf Inf

    [/CODE]



    rounds2die - How long it will take the ship to be killed by the "ideal" ship

    rounds2kill - How long it will take the ship to be kill the "ideal" ship

    rounds2killwshort - How long it will take the ship to be kill the "ideal" ship if you using Short Ranged Weapons (+50% damage)

    rounds2killwmissile - How long it will take the ship to be kill the "ideal" ship if you using Missile Weapons (+20% damage)

    rounds2killwshortmissile - How long it will take the ship to be kill the "ideal" ship if you using Short Ranged Missile Weapons (+80% damage)

    So in this system, the idea ship would actually be the 1 armor + max short ranged missile weapon ship, 2 armor + max short missiles, or the 1 def + max short ranged missiles ship.



    The problem is I don't think this system will be any more interesting or deep than the current system. It just moves around the values a little. You'd also need to put a price tag on components to evaluate the economic comparison of ships.
    0Send private message
    11 years ago
    Jul 29, 2013, 11:41:30 AM
    You are definately right with: "It isn't more deep or interesting than the current system." since that is more or less a simplified version of the current system with the main difference being, that I'm trying to meet certain expectations of what should happen in certain cases.



    It should make defenses in different quantities a viable choices and not lead to a situation where it becomes either mandatory or pointless depending on hull-size.



    I'd also rather have a system where weapons do all the same just have a different counter instead of a lot of effects disturbing their balance.



    It might also well be that my outcome expactations are not very well thought up. But this also was just an example of what I feel should always be the first step before going into concrete examples and assigning numbers.



    Also your remark of industry-efficiency is quite important. This definately has to be considered as well.
    0Send private message
    11 years ago
    Jul 29, 2013, 2:09:00 PM
    Antera wrote:
    Math is a powerful tool but there is no mathematical model for human players.




    In ES there is actually. Because battles are not tactical, we can "relativity easily" (as opposed to mathing out a human's tactical range which is nigh impossible right now) model out the various battlecard and formation options and simply determine what is the optimal state.



    Now there is a bit of poker when it comes to human vs human combat. We can model out what are the best combinations, but a human might try a riskier strategy for a bigger payoff. That still doesn't change the fact that we can model it with math and explain how those risky strategies can be very effective.
    0Send private message
    11 years ago
    Jul 29, 2013, 10:42:06 PM
    Per my above post, I don't think the card combat is as interesting as it first sounds. The gains vs potential gains vs losses math is off.

    Consider our legendary missile destroyer glass cannon vs dreadnaught combat? What cards make any sense for either side to play?
    0Send private message
    11 years ago
    Jul 29, 2013, 10:49:49 PM
    Stalker0 wrote:
    In ES there is actually. Because battles are not tactical, we can "relativity easily" (as opposed to mathing out a human's tactical range which is nigh impossible right now) model out the various battlecard and formation options and simply determine what is the optimal state.




    Hahaha, that's certain one interpretation. But what I meant when I said there's no mathematical model for human was that there's no mathematical model to predict how a given player will play the game. So there's always an element of surprise there for the devs. No matter how much they model everything out before hand, somebody is going to play it in a way that throws a wrench into their model.





    Anyways, I also find the card combat boring because its a lot like playing rock paper scissor where you can't tell what the other guy is going to do no matter how much intel you gather. That's very contrary to the spirit of strategy games where superior strategy born from excellent intel gathering is suppose to help. Still, for better or worse, the card game is a signature part of ES and separates it from a lot of other games.
    0Send private message
    11 years ago
    Jul 29, 2013, 11:08:34 PM
    The developers ought to team up with Cryptic Comet for the next expansion or Endless Space II, as that company specializes in videogames based on board game mechanics.
    0Send private message
    11 years ago
    Jul 29, 2013, 11:55:40 PM
    So I think we are having problems with crosstalk and a lack of organization. Not surprising given the time lapse nature of a forum. I propose that we start an unofficial project and use strict thread structures to keep things organized. I describe the threads and the structure below.



    1. Master Combat Mechanics Revamp Index Thread

    This thread contains summaries and links to individual proposals. Each proposal poster shall be in charge of creating and UPDATING their own post within this thread covering the progress and discussion of their own proposal. The initial post in this thread shall reference the subsequent posts for each of the proposal summaries. Comments would be encouraged to be discussed within the proposal threads, leaving the index thread clean except for proposal reference and summary posts.



    2. Combat Mechanics Revamp Proposal from [user] [anumber] : [topic]

    Each proposal shall have its own Thread. That user will be in charge of maintaining a summary post at the beginning of their Thread. After the summary post will be the first drafts of their proposal. Other users commenting would comment within that thread. The user would be responsible for maintaining a summary post of their thread in the Index Thread.





    Example:

    Thuvian makes the: Master Combat Mechanics Revamp Index Thread



    Ail has an idea about changing weapon and defense values.

    He first creates the new thread: Combat Mechanics Revamp Proposal from Ail 1: Weapon & Defense Rebalancing.

    Ail's first post in his thread summarizes the spirit of the project and current progress (first draft...).

    Ail's second post in his thread contains the body of his proposal.

    Ail updates his first post to point to the second post as the "current" status of the proposal.

    Ail then adds a post to the Master Combat Mechanics Revamp Index Thread listing his proposal thread and describing current progress.

    Ail then sends Thuvian a post to update the Index thread.

    Thuvian then adds Ail's Propsal to the initial post and references the post Ail made so users can find it without much trouble.



    Stalker0 has an idea about changing battle card balance.

    He first creates the new thread: Combat Mechanics Revamp Proposal from Stalker0 1: BattleCard Rebalancing.

    Stalker0 's first post in his thread summarizes the spirit of the project and current progress (first draft...).

    Stalker0 's second post in his thread contains the body of his proposal.

    Stalker0 updates his first post to point to the second post as the "current" status of the proposal.

    Stalker0 then adds a post to the Master Combat Mechanics Revamp Index Thread listing his proposal thread and describing current progress.

    Stalker0 then sends Thuvian a post to update the Index thread.

    Thuvian then adds Ail's Proposal to the initial post and references the post Ail made so users can find it without much trouble.



    Stalker0 has a second idea that is separate from his first about changing weapon ranges.

    He first creates the new thread: Combat Mechanics Revamp Proposal from Stalker0 2: Weapon Ranges.

    Stalker0 's first post in his thread summarizes the spirit of the project and current progress (first draft...).

    Stalker0 's second post in his thread contains the body of his proposal.

    Stalker0 updates his first post to point to the second post as the "current" status of the proposal.

    Stalker0 then adds a post to the Master Combat Mechanics Revamp Index Thread listing his proposal thread and describing current progress.

    Stalker0 then sends Thuvian a post to update the Index thread.

    Thuvian then adds Ail's Proposal to the initial post and references the post Ail made so users can find it without much trouble.



    etc.





    The user names are just examples. I don't care who creates the index thread. They just need to be responsible and keep up with things. The system does seem a little complex, but it really isn't. You just create your own thread, don't spread your idea across several threads, and then post a single reference to the index thread. This will keep our ideas focused in one place and make it easier for people to keep up with the different ideas. It will also decrease the amount of off-topic or side discussions that forum, making proposal much more consistent.



    The thread titles need to be consistent. It would be nice if we could group them by topic, but that's going to be awkward since many of the topics will overlap. Grouping them by user name and proposal number per user seems straightforward. We could also create an proposal number system that is overall instead of by person. So the first proposal would be 1, the second 2, etc. We'd have the Index thread keep the numbers straight.



    Looking at my love-hate affair with glass cannons you can see how what we have is a problem. People continually bring up "Defenses are too good" without reading any of the many "suggestions" that defenses aren't good enough posts that I've created backed by math. At least this way we can just answer with a link to a single Glass Cannons work better thread. We could do that now, except this thread is filled with all sorts of things, many of which don't relate to the initial idea.



    If at least 2-3 people are willing to work with me to create this, I think we should go for that plan, and adjust our posting to fit it. I don't want to make anyone do this, but I do think it would create a system that would work much better than our current method of posting where-ever. Additionally, does anyone want to create the Index thread, I'll do it if no one else wants to, but I'm lazy.



    State of Affairs summary:

    I'm currently aware of 3 or 4 balance suggestions in varying states of organization. That would mean we'd already have three proposals to restart and index. Note that we can have proposals that suggest different outcomes, that's the entire idea behind the proposals. We put together some ideas and see which ones work better. We should also get the guy who made the big rebalance mod to contribute some of his thoughts. Given how much work he has already done it seems reasonable that he already has a lot to say about the subject.
    0Send private message
    11 years ago
    Jul 30, 2013, 7:14:11 AM
    The guy with that Rebalance Mod is "Foraven". You might want to send him the link to this thread via PM.
    0Send private message
    11 years ago
    Jul 31, 2013, 2:05:07 AM
    One of the problems I've been having in formulating a balance system that allows for a purpose, but not overwhelming use of Glass Cannons, is the difference between long, medium, and short range weapons.



    As they stand, their names are actually a huge lie! A "long" range weapon is slower than a "short" range weapon. The Long Range Missile gets to its target in 4 rounds, whereas the Short Range Missile only takes 1 round to hit. That's weird. My first attempt to fix this was to simply make all weapon speeds the same (fast). This makes combat a little less flashy, where rounds 1 is the flashiest followed by round 3. Rounds 2 & 4 are rather dull. Trying to fix this by changing how long a reload takes also resulted in an unsatisfactory response.



    That's when I (finally) realized what "long range" meant. Let's ignore titles and just talk about characteristics. Missile A is slow, accurate, does a lot of damage, and you only fire one at a time. Missile B is fast, inaccurate, does little damage, but you fire lots of them. So in the first category we have guided missiles (AGM 114-Hellfire), in the second we have dumbfire rockets (e.g., Hydra 70) (note the example weapons speeds aren't correct in this analogy, but you get the drift.).



    So that means that perhaps thinking of Long, Medium, and Short Range Combat phases is also not the way forward.
    0Send private message
    ?

    Click here to login

    Reply
    Comment