Pages: 1 [2] 3 :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |
Akyla
Bears Inc Violent-Tendencies
|
Posted - 2008.11.05 13:55:00 -
[31]
Originally by: dojocan81
Damage = Base_Damage * MIN(MIN(sig/Er,1) , (Ev/Er * sig/vel)^(k * drf) )
Er = Explosion Radius of missile
Er = Base_Er*(1-gmp_lvl*0,05)
Ev = Explosion Velocity of missile
Ev = Base_Ev*(1+0,1*tmp_lvl)
Hmm... I guess that's quite obvious. I'm slow today... got a cold :S ________________________________ All your honey are belong to us! |
Bladen Kerst
Caldari Locked and Loaded
|
Posted - 2008.11.12 08:24:00 -
[32]
bump
|
Stafen
Killer Koalas
|
Posted - 2008.11.12 16:52:00 -
[33]
I have been testing the six types of T1 missiles on Sisi and after the testing I have to update the formula slightly.
The update explains why the cruise missiles (and other high Damage Reduction Factor missiles) don't quite fit in cleanly.
I first tested the 6 T1 missile types (light, rocket, heavy, ham, cruise and torp) and fitted the data to the equation:
Damage = Base_Damage * MIN(MIN(sig/Er,1) , (Ev/Er * sig/vel)^c )
where 'c' is a constant which was different for each missile.
The all fitted the equation nicely but with different values of c dependent on the drf.
missile / drf / c light 2.8 0.603838 rocket 3 0.639119 heavy 3.2 0.679911 ham 4.5 0.881599 cruise 4.5 0.879992 torp 5 0.939506
The numbers seem to fit the equation: c = K * log(drf)
as you can see from this graph
Thus my best guess for the formula is now:
Damage = Base_Damage * MIN(MIN(sig/Er,1) , (Ev/Er * sig/vel)^(0.586 * log(drf))
(log is the natural logarithm)
There needs to be more testing with T2 missiles (which I cannot use) to get more values for drf to compare against.
Also the number 0.586 is only my best guess.
Thank you everyone who reads this and for those of you who have tested it. I will update the first post.
|
Mikhale Romanov
Black Hats Delta
|
Posted - 2008.11.12 17:35:00 -
[34]
hehe.. Now.. just out of sheer curiosity.. would anyone say that someone at CCP forgot a few 0's in their math?~ Just a thought ZOMG Communism! |
Gypsio III
Dirty Filthy Perverts
|
Posted - 2008.11.12 17:35:00 -
[35]
Quote: missile / drf / c light 2.8 0.603838 rocket 3 0.639119 heavy 3.2 0.679911 ham 4.5 0.881599 cruise 4.5 0.879992 torp 5 0.939506
The numbers seem to fit the equation: c = K * log(drf)
Thanks, Stafen
Just to clarify, the log function is the natural logarithm, to the base e, often denoted as ln.
|
Stafen
Killer Koalas
|
Posted - 2008.11.12 17:41:00 -
[36]
Originally by: Gypsio III
Quote: missile / drf / c light 2.8 0.603838 rocket 3 0.639119 heavy 3.2 0.679911 ham 4.5 0.881599 cruise 4.5 0.879992 torp 5 0.939506
The numbers seem to fit the equation: c = K * log(drf)
Thanks, Stafen
Just to clarify, the log function is the natural logarithm, to the base e, often denoted as ln.
Correct, but I use matlab (octave) for my maths and I am a mathematician so logarithm are normally base e :-) My pocket calculator uses ln though.
|
Gypsio III
Dirty Filthy Perverts
|
Posted - 2008.11.12 17:49:00 -
[37]
Just running the modified formula though my spreadsheet. The difference relative to the previous iteration appears to be minimal - except for Rage and Fury, which do somewhat more damage then previously modelled in absurd situations such shooting Rage torps against an interceptor.
|
dojocan81
|
Posted - 2008.11.12 17:53:00 -
[38]
good work on that!!!
posting some graphs later
|
Celeritas 5k
Caldari Perkone
|
Posted - 2008.11.12 19:22:00 -
[39]
Thanks!
It might be clearer to write it as (Ev/vel * sig/Er), rather than (Ev/Er * sig/vel). Makes the relationships between the numbers more obvious... that's just me though :D - Always be Happy, Never be satisfied. |
Wannabehero
Caldari Absolutely No Retreat
|
Posted - 2008.11.12 19:41:00 -
[40]
Stafen is the one on the right --
Don't harsh my mellow |
|
Leeluvv
Federation of Freedom Fighters Executive Outcomes
|
Posted - 2008.11.12 22:59:00 -
[41]
Edited by: Leeluvv on 12/11/2008 23:03:12 Edited by: Leeluvv on 12/11/2008 22:59:40 I've tried to simplify the formula into english so luddites like me can understand what it means in game terms (it isn't perfect as the last bit of the formula isn't a value that is easy to describe verbally).
If the target ship's Signal Radius is less than missile's Explosion Radius, damage is reduced by the ratio of them, otherwise the damage is reduced by the ratio of the missile's Explosion Velocity to the ship's Velocity to the power of (constant x damage reduction factor)
It may be wrong, but I needed a description to be able to relate the formula to what this meant in the game. I'll post some of the working to explain how I came up with it:
Original formula: Damage multiplier = MIN(MIN(sig/Er,1) , (Ev/Er * sig/vel)^c )
but, (Ev/Er * sig/vel) = (Ev/vel * sig/Er)
Formula can now be expressed as: Damage multiplier = MIN(MIN(sig/Er,1) , (Ev/vel * sig/Er)^c )
sig/Er is now common to both values being compared, so the formula can be expressed as: Damage multiplier = MIN(sig/Er,1) * (Ev/vel)^c
Lee == Sig to follow |
Kaven Kantrix
Two Brothers Mining Corp. Kraftwerk.
|
Posted - 2008.11.13 01:48:00 -
[42]
Looking at your spreadsheet, assuming I am reading it right (since I cant do the formula in my head), if you want to increase damage on a target, even if the target is going very very fast, you can do so simply by increasing it's signature radius.
Or, in other words, applying a target painter to a fast moving target will increase damage no matter what. Even if my missile has an explosion velocity of 80, and my target has a velocity of 1000, my target painting will increase damage applied.
Is that right?
|
Gypsio III
Dirty Filthy Perverts
|
Posted - 2008.11.13 01:54:00 -
[43]
Yes.
To do full damage, your target must have a sig radius greater than the explosion radius of your missile. This is the same as the old mechanics.
However, if your target is going faster than the explosion velocity of your missile, then you can still deal full damage if the target's sig is sufficiently greater than your missile's explosion radius.
This can be achieved increasing the target's sig using Painters or the 500% self-painting effect of a MWD, or by decreasing the explosion radius of your missile, such as by using Rigour rigs, Crash or by training Guided Missile Precision.
|
Stafen
Killer Koalas
|
Posted - 2008.11.13 02:01:00 -
[44]
Originally by: Kaven Kantrix Looking at your spreadsheet, assuming I am reading it right (since I cant do the formula in my head), if you want to increase damage on a target, even if the target is going very very fast, you can do so simply by increasing it's signature radius.
Or, in other words, applying a target painter to a fast moving target will increase damage no matter what. Even if my missile has an explosion velocity of 80, and my target has a velocity of 1000, my target painting will increase damage applied.
Is that right?
That is correct, if you are not getting full damage from your missiles a target painter will always increase your damage (by how much is the question) (using the old formula pre-QR that was not the case when speed tanking)
As % damage received is the minimum of the following values: 1 s/Er (Ev/Er * s/v)^c
1 is full damage (so TP no help), other two numbers both get larger if the target signature increases.
The spreadsheet I linked in the first post uses an old formula which has been outdated but it is almost correct.
|
Allen Ramses
Caldari Typo Corp
|
Posted - 2008.11.13 02:10:00 -
[45]
Edited by: Allen Ramses on 13/11/2008 02:12:02
Originally by: Leeluvv Damage multiplier = MIN(sig/Er,(Ev/vel)^c ,1) <= I don't know if that's a correct use of MIN, as I didn't even know what it was until I read this thread and hit F1 in Excel.
Yes, that is the correct usage of MIN, but that is not the correct equation. Basically, to break it down, it is this.
MIN( <---Hello World.
Er/Sig; <---Damage multiplier directly proportional to signature radius
((Ev/Er) * (Sig/Vel)) <---Stage 1 variable reduction. The product of two ratios. This sets the reference damage multiplier.
^(0.586 * log(drf)); <---Stage 2 variable reduction. Uses exponential multiplication to alter the reference damage multiplier. The lower the number, the closer to 1 the above number will effectively be. This is like a falloff.
1) <---Damage multiplier is not to exceed one. Goodbye World.
Anyway, that should clear that up. However, one thing I am unclear on is the logarithmic function. To use LN(drf) gives me a much different result from using LOG(drf). Which is correct?
EDIT: Damn forum colors... ____________________ Pimped out Raven to run level 4 missions quickly: 210 Mil ISK. Realizing your 120 Mil ISK Drake gets the job done faster: Priceless. Made a reality by speed and missile nerf. |
dojocan81
|
Posted - 2008.11.13 06:31:00 -
[46]
Originally by: Allen Ramses
Anyway, that should clear that up. However, one thing I am unclear on is the logarithmic function. To use LN(drf) gives me a much different result from using LOG(drf). Which is correct? EDIT: Damn forum colors...
LN(drf)is the correct one
|
Skie Ankora
|
Posted - 2008.11.13 18:14:00 -
[47]
I built a quick spreadsheet to draw graphs comparing missile damage against target speed for both current and pre-quantum rise missile damage formulas (using the formula in this thread and one from 2005 I found here: http://oldforums.eveonline.com/?a=topic&threadID=198863).
There is a dropdown where you can select the missile type and the target ship class and whether or not they have a MWD active.
It is very quick and dirty at the moment, and may well contain errors, but I thought I would post a link to it here in case anyone else wants to take a look or do further work on it.
Link to spreadsheet: http://eve-files.com/dl/178366
|
Viktor Del'Grande
|
Posted - 2008.11.13 19:10:00 -
[48]
Hi guys,
today i made some test on sigularity and collected some data. The collected curves shows that Stafen's formula is right.
Here is my Scilab file and a screenshot of the diagram:
Scilab File Diagram
The Data is not DPS wise, its the single hit at 0% Resists. The data got collected at a Ishkur. The Resists of the AF and the fitting of the firing raven is documentated in the scilab file.
|
Stafen
Killer Koalas
|
Posted - 2008.11.13 19:12:00 -
[49]
Originally by: Skie Ankora I built a quick spreadsheet to draw graphs comparing missile damage against target speed for both current and pre-quantum rise missile damage formulas
Nice spreadsheet, a nice quick way to test things. (I suck at making spreadsheets)
Also I thought I should post data to backup my claims.
First my data, QR_missile_data.png
As you can see out of my tests of all T1 missiles the biggest error I get is 0.82%. Would like to know how it works with T2 missiles.
The data plotted to find the constant in the formula: QR_missile_const.png
Also some interesting maths regarding Target painters:
1) All damage increases is capped by by the base missile of the missile of course.
2) If the target is stationary or not fast enough to speed tank the damage increase in % is the same as the signature radius bonus. I.E. unbonused T1 painter gives 25% damage increased.
3) If the target is in the speed tanking range (which is most of the time) the damage increase for a missile is: (TP bonus)^(0.586*log(drf))
so for a T1 TP and cruise missile: 1.25^(0.586*log(4.5)) = 1.2173 so 22%
for T1 TP and torps: 1.25^(0.586*log(5)) = 1.2342 so 23%
nice and simple and tested in-game :-)
|
Javius Rong
Caldari Sigillum Militum Xpisti
|
Posted - 2008.11.13 19:19:00 -
[50]
Thanks for all the hard work. Upon reviewing the formula and the shape of the damage reduction factor pre and post, I am very surprised at how fast the new one falls off (width and slope). I would have expected a wider shape to the curve with less slope. Does this curve look similar to a rail gun tracking fall off with transversal speed?
|
|
Viktor Del'Grande
|
Posted - 2008.11.13 22:17:00 -
[51]
I did a update and now you see Stafens theoretical data and my collected data in two diagrams and they match really good
New Scilab File New Diagram
|
Diziet Montoya
|
Posted - 2008.11.13 22:41:00 -
[52]
When I click "Show Info" on loaded cruise missile (fitting screen) it says explosion velocity 96.6 m/s. When I do same for a torp it says 99.4 m/s. Surely the figure for cruise should be higher than torps?
|
Kaven Kantrix
Two Brothers Mining Corp. Kraftwerk.
|
Posted - 2008.11.14 00:43:00 -
[53]
Originally by: Stafen 1) All damage increases is capped by by the base missile of the missile of course.
Uh...so I am guessing the bolded text should be "by the base damage"?
|
Stafen
Killer Koalas
|
Posted - 2008.11.14 00:48:00 -
[54]
Originally by: Kaven Kantrix
Originally by: Stafen 1) All damage increases is capped by by the base missile of the missile of course.
Uh...so I am guessing the bolded text should be "by the base damage"?
Yep, thanks, did not proof read it carefully enough.
|
Stafen
Killer Koalas
|
Posted - 2008.11.14 00:57:00 -
[55]
Originally by: Viktor Del'Grande I did a update and now you see Stafens theoretical data and my collected data in two diagrams and they match really good
New Scilab File New Diagram
Nice program scilab, but I have not much experience using it.
I am guessing most of the errors between theoretical and measured damage is due to the high resists of the target (which makes the damage inflicted small thus rounding plays a bigger role) and the complexity of the setup of the attacker and target.
But I like the test as it tries a large amount of modules which effect missile damage thus showing how everything works together.
|
Skie Ankora
|
Posted - 2008.11.14 01:40:00 -
[56]
Originally by: Javius Rong
I am very surprised at how fast the new one falls off (width and slope). I would have expected a wider shape to the curve with less slope.
As am I - it seems like in most situations missile damage will be significantly less than it was pre-nerf even when using the correct weapon type.
For example using heavy missiles against a cruiser sized target, the damage is 35% less than pre-nerf levels at speeds of 200m/s (which can be achieved easily without using an afterburner).
Fitting an afterburner helps reduce damage as expected, but not as much as it should compared to a ship moving unaided at full speed. AB typically reduces damage by a further 15% or so on most ships, so 50% reduction in total.
The exception is when the target activates the MWD, when the damage is similar or even more than pre-speed nerf levels, as expected.
Originally by: Javius Rong
Does this curve look similar to a rail gun tracking fall off with transversal speed?
That's an interesting point I looked into it and it turns out that the relationship is actually linear. If you double transversal, chance to hit reduces by a factor of 4, and so on.
So no the relationship is very different - from looking at the formula these changes mean that missiles are nerfed for about a third less damage unless the target is using a MWD.
|
Allen Ramses
Caldari Typo Corp
|
Posted - 2008.11.14 02:30:00 -
[57]
Originally by: Skie Ankora As am I - it seems like in most situations missile damage will be significantly less than it was pre-nerf even when using the correct weapon type.
That's quite the understatement. As it is now, assuming max skills, a torpedo will deal only 40% damage to a typhoon going 162m/s. No, not after resists. ____________________ Pimped out Raven to run level 4 missions quickly: 210 Mil ISK. Realizing your 120 Mil ISK Drake gets the job done faster: Priceless. Made a reality by speed and missile nerf. |
Viktor Del'Grande
|
Posted - 2008.11.14 07:06:00 -
[58]
Originally by: Stafen
Originally by: Viktor Del'Grande I did a update and now you see Stafens theoretical data and my collected data in two diagrams and they match really good
New Scilab File New Diagram
Nice program scilab, but I have not much experience using it.
I am guessing most of the errors between theoretical and measured damage is due to the high resists of the target (which makes the damage inflicted small thus rounding plays a bigger role) and the complexity of the setup of the attacker and target.
But I like the test as it tries a large amount of modules which effect missile damage thus showing how everything works together.
As i saw the data, this was my tought too. Today i will do the more "important" test against speed tanked cruiser. Maybe i can get a thired char which remote reps the "poor target"
|
Murina
The Scope
|
Posted - 2008.11.14 08:15:00 -
[59]
Gunships have a min/max range (optimal+fallof) lets say 0-30km, now in that 30km they need to use their weakest long range ammo and even that suffers from dmg reduction at some of the 0-30km ranges.
Or if they want to use the highest dmg ammo for the range the target ship is at they need to swap ammo every time the target ship moves to another range.
And on top of that theirs the whole transversal issue.
Now missiles may receive a dmg reduction due to speed/sig but they do not get one in regards to range or transversal and not needing to swap ammo or just use the lowest dmg type because of a range issue is a great benefit.
Both systems have down sides deal with it.
POST NERF PVP SKILLS: "shall we engage?" "hmmm how many ships do they have?" "more than us" "lets not bother then" "WOW great job FC!!!!" "................. |
Skie Ankora
|
Posted - 2008.11.14 08:52:00 -
[60]
I updated my post above ^^^ for clarity, and to add some information I omitted about the negative effects of shooting missiles at a smaller ship class (e.g. cruise missiles against frigates), which is now significantly reduced on pre-patch levels even if the target activates a MWD.
|
|
|
|
|
Pages: 1 [2] 3 :: one page |
First page | Previous page | Next page | Last page |