| Pages: 1 [2] :: one page |
| Author |
Thread Statistics | Show CCP posts - 0 post(s) |
ELECTR0FREAK
   |
Posted - 2005.07.07 12:34:00 -
[31]
Edited by: ELECTR0FREAK on 07/07/2005 12:38:40
Light, explosion velocity 1300m/s with a missile explosion radius of 42m and a base damage of 95.8.
Heavy, explosion velocity 975m/s with a missile explosion radius of 127m and a base damage of 190.8.
Everything was fired from a Cerberus with no damage mods at varying frigates. All used MWDs and thus their signature radius in all velocity tests was greater than the explosion radius of the missile. This creates an initial multiplier of 1 for all missile damage hits, and thus I isolated testing to just the velocity and not signature radius effects. It's possible I missed something, but I don't think so.
Differing frigs were needed to achieve different stable velocities, but I made sure all met the criteria of having a sig radius > the explosion radius that I was firing. I also used only EM missiles on the shields of my targets, so no resistances were involved. Almost all values were recorded after the target reached a stable orbit velocity, except for very high speeds (4km+) which had to be achieved via a fly-by scenario.
This information (and more) is contained in the .XLS file, in the 2 columns farthest right.
I've got to go get some sleep (8:30 AM here in the states and I have work in several hours), so good luck everyone, you have my best wishes. 
|
Vishnej
|
Posted - 2005.07.07 12:38:00 -
[32]
Edited by: Vishnej on 07/07/2005 12:40:29
Missile velocity, not explosion velocity - I suspected that it plays a part due to the fact that the bottom of the light missile graph seems to trend towards x*((speed-4000)/4000).
With colth's work, I'm not so sure though.
Torps would be wonderful to work with to find out, after determining the Torp Constant.
The constants look approx. proportional to explosion speed - 1300/2.125=6112, 975/1.525=639.
|
ELECTR0FREAK
|
Posted - 2005.07.07 12:41:00 -
[33]
Oh, uhm, all guided missiles were traveling at a speed of 5,675 m/s, largely in part to the missile speed bonus of my Cerberus.
I don't see how missile velocity itself can effect damage, but I suppose it IS possible. Wouldn't that give Ravens, Cerbs, Caracals etc a pretty hefty advantage though?
With that, I'm out for now... good luck!
|
MrMorph
|
Posted - 2005.07.07 12:42:00 -
[34]
Might i remind you guys that missiles and modules that affect speed/radius is borked ?
Heavies does the same damage as a torp atm. (30 dmg to a frig that is web'ed and targetpainted)
----------------------------------------------
Trishys cookies they are !
|
ELECTR0FREAK
|
Posted - 2005.07.07 12:51:00 -
[35]
Well part of that has to do with the fact that a torp has almost 3x the explosion radius of a heavy missile, (nerfing it's damage versus small targets by 37%) but at the same time does 3x the damage, which counters this fact.
As for the rumors that target painters have been doing screwy things to missile damages, I can't say. They worked just fine at increasing damage against small ships when using larger missiles when I was testing them. In fact, a Target Painter I improves missile damage by the same amount in paints for, 25%, provided the target's sig radius remains below the missile's explosion radius.
Alright. Last time, I'm off. 
|
elFarto
|
Posted - 2005.07.07 13:25:00 -
[36]
 Originally by: Colthor
Const depends on the type of missile:
Light missile: const = 2.125 (approx.)
Heavy missile: const = 1.525 (approx.)
Those numbers are awfully close to missle damage/explosion radius.
Regards
elFarto
|
Petrus Illyusanov
|
Posted - 2005.07.07 15:57:00 -
[37]
for the missile damage formula based on target velocity that appears to be a guassian curve that you can fit with high accuracy with the formula:
DAMAGE_AT_VELOCITY_X = BASE_DAMAGE*EXP(-1*((1300-X)/1300)^2)
Or a guassian with the mean and standard deviation equal to the missiles blast velosity. This is with the assumption there is some code in there that ignores velocities below 1300 m/s as irrelevant. It isn't exact but it is within a point or two and is likely to be tolarable.
|
ELECTR0FREAK
|
Posted - 2005.07.08 06:26:00 -
[38]
Edited by: ELECTR0FREAK on 08/07/2005 06:29:18
I've FIGURED IT OUT! At least, the formula I'm using supports all of my data within a fraction of a decimal place in damage, which I pin down to rounding and error in data. So, without further ado, here is my formula:
First, there are conditions that the formula must adhere to. Keep these in mind when making your final calculation.
If Target Signature Radius >= Missile Explosion Radius, ( Target Signature Radius / Missile Explosion Radius ) = 1.0
If Target Velocity <= Missile Explosion Velocity, ( Target Velocity - Missile Explosion Velocity ) = 0.0
Missile Damage * (Target Signature Radius / Missile Explosion Radius) * EXP( -1 * (Target Velocity - Missile Explosion Velocity )^2 / (1500^2) ) = Final Missile Damage
This is definately too complicated to do casually, but I'm sure Mr. M will be happy to incorporate this equation into his missile damage calculator at www.evegeek.com
Anyhow, in conclusion, I'd like to thank everyone on this thread who's helped thus far (especially Petrus for getting me onto the Gaussian Curve route), and I'd also like to thank Badri and CDLPeacemaker for suffering through all my testing.
I also take donations in the form of isk, don't forget the zero in my name... 
|
The Pacman
|
Posted - 2005.07.08 08:25:00 -
[39]
hi
just a question?
and you think this equation can be handled by the server in a fleetbattle or so? When a lot of missiles are fired and a lot of damage has to be computed? 2 potencials and an exp function? Thats definitly to much for a server which has to handle some hundred or even tousands of this computations in a minute.
I don't want to say that your work is bad but i think that can't be the formula at all. The first error i've seen in your data is that you already using the new skills. This may be a reason your results doesn't match with the simple formulas wrote down in this thread before. The patch isn't 3 days old and it may be that there is somehow a bug in the new missile skills. So i just can advise - use a character with none of the new skills for testing. I've done so and also figured out a possible formula which is a much easier computation for a program. The formula i think is the follows:
missile damage = base_missile_dmg * min(min(explosion_speed/target_speed;1.2)*min(target_sig_radius/explosion_radius;1.2);1.2);
As you can see it is mainly based on the formula writen by Mr M. It just uses other upper borders for the factors. The min(..) are normal min-functions like in C. It takes the smaller of the both values provided. The upper border for the factor of the damage is 1.2 and not as writen in this thread before 1.0. Why? Just because the missile dmgs before & after patch are in this ratio togehter. Example: Take a ship with a sig radius > 100 (cruiser for example) and set it's speed to 0. Fire with a light missile on it. You will see that your missile will make 90 dmg without resistences. But with none of the new skills the base damage of light missiles is 75. 90/75 = 1.2 .
This is tested by me with a kestrel and a caracal. But only up to speed around 1000m/sec. As i can see the major speed factor decrease will come into there and above so i've to test it with someone who can fly such speeds without an mwd. This will be done at the weekend i think. Maybe some of u guys can test it too and tell me your results here.
thx & bye
The Pacman
-- SteelVipers --
|
Petrus Illyusanov
|
Posted - 2005.07.08 09:15:00 -
[40]
I would just like to say that frankly the data that was given in the zip file for missile damage as a function of target velocity is a guassian distribution. From what I can see the game uses a guassian for the effect of gun fire (chance to hit over optimum range is likely a guassian with width equal to fall off) so why not for a missile hit? At least what I saw on someones web site on eve mechanics looked pretty much guassian to me.
Expontential functions are standard mathematical functions and not exactly computationally excessive. Also under most normal conditions a simple "if Target_velocity < Missile_blast_velocity then 1" gets rid of the need to calculate it anyway.
So only in intercerptor battles is the formula likely to be used anyway. Frigates with MWD barely exceed the blast velocity.
But unless the developers wish to comment in the end we will never settle the question.
|
ELECTR0FREAK
|
Posted - 2005.07.08 09:43:00 -
[41]
In response to Pacman, the formula works flawlessly. It matches every point of my data within a fraction of a point of damage.
So, it works, without a doubt. And I have to agree with Petrus, I do a fair bit of programming myself as a Comp Sci student and most low-power exponential functions are pretty easy to process. Squaring something is essentially just multiplying it by itself. So if you break it down, all the formula is is a series of multiplication and division problems.
|
KIAEddZ
|
Posted - 2005.07.08 10:11:00 -
[42]
SO!!!!!
How much actual difference will skills make in a Battle between:
Frig vs Frig (rockets and std missiles)
Cruiser vs Frig (Heavy Missiles)
Cruiser vs Cruiser (Heavy Missiles)
Battleship vs Frigate (Cruise Missiles)
Battleship vs Cruiser (Cruise Missiles)
Battleship vs Battleship (Cruise Missiles)
http://oldforums.eveonline.com/?a=topic&threadID=176347
www.kia-corp.co.uk/killboard
CEO of KIA Corp - Been doing it for the Laydeez since 1993, now we i |
ELECTR0FREAK
|
Posted - 2005.07.08 10:51:00 -
[43]
Edited by: ELECTR0FREAK on 08/07/2005 10:52:11
A good question, and doubtless is something all missile users have foremost in their minds.
I can tell you that training Guided Missile Precision, because the part of the equation that handles that is a fairly straightforward division problem, should give you about as much of a damage increase versus targets smaller than your missile explosion radius equal to the 5% bonus. This is a bit confusing so here's an example:
Heavy Missile being used against an Interceptor with 26 signature Radius (ie Crow, Inty lvl 4).
Guided Missile Precision lvl 0 (150m explosion radius):
26/150 = .1733 damage penalty. Essentially, the missile hits for 17% of normal damage, assuming 180 (Heavy Missiles IV) = 31.2 damage.
Guided Missile Precision lvl 5 (112.5 explosion radius):
26/112.5 = .2311 or 23% reduction in damage. Again assuming 180 damage, this is 41.6.
Since 31.2/41.6 = .75, 31.2 is 75% as much damage as 41.6, meaning that essentially, there was a 25% reduction in the damage penalty. This means that for every level of Guided Missile Precision you train, you are effectively improving your large missile damage versus small targets by 5%.
Now, since training Target Navigation Prediction basically makes your missiles explode faster, it requires a target to be going faster before your missiles suffer any kind of damage penalty. With Target Navigation Prediction 0, if an enemy is going faster than 750 m/s, your missiles begin to hit for reduced damage. Train it to lvl 5, and the target has to be going faster than 1125 m/s before your missiles take any damage penalties. The slope of both graphs is exactly the same, just by training the skill you are shifting it over. Here's another example:
Heavy Missile being used against a MWDing Interceptor (thus it's signature radius is > missile explosion velocity and doesn't play into this arguement). Interceptor is going 3000 m/s.
Target Navigation Prediction lvl 0 (750 m/s explosion velocity):
According to my graphing calculator, resulting damage will be 20.1 damage (rounded down).
Target Navigation Predition lvl 5 (1125 m/s explosion velocity):
According to the calculator, damage will be 40.0 damage (rounded up).
Despite the rounding differences, it's fairly easy to see that 20.1 is approximatly 50% of 40, so training the skill to level 5 reduced the damage penalty by 50%. Again, this coincides with the fact that the skill provides a 10% bonus per level trained. Now, keep in mind, Rockets and Torpedos do NOT benefit from training this skill, only Light, Heavy and Cruise missiles will see this improvement.
So, as you can see, training the skills are definately worth the time. It won't let you hit small fast ships with big missiles like you did before, but it decreases the damage penalty they suffer by a pretty big degree.
Edit- 7am... perhaps time to get some sleep.  And I've gotta go to work today...
|
KIAEddZ
|
Posted - 2005.07.08 10:55:00 -
[44]
Without a shadow of a doubt Electro, that is the most hope filled post I have seen yet.
Someone needs to spend the time, and do the calcs for the various missiles vs the various size and types of ship.
These posts will give people hope that its not all doom and gloom.
Pld m8.
http://oldforums.eveonline.com/?a=topic&threadID=176347
www.kia-corp.co.uk/killboard
CEO of KIA Corp - Been doing it for the Laydeez since 1993, now we i |
ELECTR0FREAK
|
Posted - 2005.07.08 11:03:00 -
[45]
Edited by: ELECTR0FREAK on 08/07/2005 11:08:46
Well bud, I've gotta head to bed (7am here and I've got work today!) but I'll put it on my to-do list for this evening, provided somebody doesn't beat me to it.
Edit- hell, I'll start a new topic for that post as probably most of this forum's regulars haven't bothered to keep tabs of this thread. I'd like everyone to get a good grasp of how it all works. Maybe later tonight after work I can add to it with some pretty graphs made in Excel.
|
The Pacman
|
Posted - 2005.07.08 11:10:00 -
[46]
and hi again
Electr0Freak: of course it matches your data - it's simply logical - it's build up on your data. But like i said before. Your data has the new missile skills involved. This is a potential error source, you have to know. Try to make a new test run with a character without the new missile skills learned. And then tell me if your formula still works ...
and now an example why i think it doesn't work...
target ship: kestrel - signature radius: 47m - speed: 0 m/sec
missile:
light emp (dont know the name atm) - signature radius: 50m expl. velocity: 1000m
Quote:
If Target Signature Radius >= Missile Explosion Radius, then ( Target Signature Radius / Missile Explosion Radius ) = 1.0
Quote:
If Target Velocity <= Missile Explosion Velocity, then ( Target Velocity - Missile Explosion Velocity ) = 0.0
Quote:
Missile Damage * (Target Signature Radius / Missile Explosion Radius) * e^( -1 * (Target Velocity - Missile Explosion Velocity )^2 / (1500^2) ) = Final Missile Damage
using this formulas for the facts writen above:
final missile damage = 70.5 hp without any resistences - but thats the theory
real missile damage done by test: 84.6 hp without any resistences
ok now it's ur time to tell me what i've calculated wrong ....
bye
The Pacman
p.s.: i really don't want to **** off your work - this thread gaves me the final hint which leads me to the formula i posted before - so i have to say thanks for that. I only want to know the right formula too.
p.p.s: i've studied computational sciences too and i know what i say when i say that this formula is in a computational view to complex to be computed for n-thousand times in a minute
|
ELECTR0FREAK
|
Posted - 2005.07.08 11:21:00 -
[47]
Edited by: ELECTR0FREAK on 08/07/2005 11:26:52
Uhm... buddy, you've got Standard Missiles IV on that character. Base damage is 90, multiply that by 47/50, result is 84.6.
I appreciate you checking my work and all but believe me, I wouldn't have posted my formula if there was any doubt in my mind that it was incorrect. I wish I could show you my graphing calculator right now. All my points of data that I have stat plotted are run right through with 2 curves, one right through every single point of my frigate data, and one right through every single point of my cruiser data.
I've REALLY gotta get some sleep, but I'll be happy to make some new Excel graphs tomorrow showing this.
|
Petrus Illyusanov
|
Posted - 2005.07.08 11:41:00 -
[48]
Originally by: The Pacman
using this formulas for the facts writen above:
final missile damage = 70.5 hp without any resistences - but thats the theory
real missile damage done by test: 84.6 hp without any resistences
ok now it's ur time to tell me what i've calculated wrong ....
bye
The Pacman
p.s.: i really don't want to **** off your work - this thread gaves me the final hint which leads me to the formula i posted before - so i have to say thanks for that. I only want to know the right formula too.
p.p.s: i've studied computational sciences too and i know what i say when i say that this formula is in a computational view to complex to be computed for n-thousand times in a minute
*sigh* 75 is th base damage modified by standard missile skill level that in Electro's data gives a 20% increase in damage and that now gives a new base damage of 90 and then reduce it by 6% due to signature and you get the damage of 84.6 which is what you saw. Can I assume you have standard missile 4?
As for computationally intensive give me a break.
|
Hoshi
|
Posted - 2005.07.09 19:50:00 -
[49]
As to the question about if the server can handle the equations. Should not be a problem as if it becomes a bottle neck you can always precompute most of it and do a quick table lookup instead. This is how most other games solve it.
|
GSBarlan
|
Posted - 2005.07.31 03:04:00 -
[50]
 GREAT JOB ELECTRO!!!! GFJ!!!!!!!!!!
|
ELECTR0FREAK
|
Posted - 2005.07.31 03:05:00 -
[51]
Whats with people ressurecting my missile threads?
-Electrofreak
Discoverer of the Missile Damage Formula |
| |
|
| Pages: 1 [2] :: one page |
| First page | Previous page | Next page | Last page |