Weaponsmith Archive

Hello,

I will use these posts to sort of log my formula and ideas mostly, so lots of
posts by me, but any fixes and ideas welcome.

I am a master pikeman / master swords so that is the info I worked on.

How you fight in game:

With meaningful special action costs you can no longer spam the best move (tm)
as you will run out of action.

I will ignore state attacks in this dicussion as they complicate things, and
most of the time just add an additional damage multiplier.

Also I will assume that you cannot spam your most damage causing attack
continously as you would run out of action (if you can then just do that)

Now with my template I do auto attack "melee hit" and then spam "improved head
hit" as much as my action will allow.

If I just do "melee hit" my action will never run out, (it will with "melee
strike")

Data:

Default rate of action recharge on a Zaberk with no action buffs:
about 2.75% points a second (I will call this rps - regen per second)
( I just opened charater sheet and used a stopwatch - if anyone wants to get
more accurate numbers you are welcome )

Now from:
Sword Specials

you can see only melee hit will allow you to attack continously with any
reasonable weapon speed. (about 2.6 modified for my powerhammers)

Melee Strike would burn 10/2.6 = 3.84 action a sec, more than my recharge.

( yes lower sac weapons might allow it, I found if you take that post's action
cost number * sac * (18/(120*15)) I get the action cost for my setup - but all
my weapons give the above result )

In a fight I alternate with a ratio between two hits melle hit and impr. Head
hit.

(note: there is a cooldown for head hit but it is short enough that doing
melee hits during cooldown still allows me to drain all my action so I am
fighting to my action regen capacity)

Now looking at:
Damage output and weapon Stats

You see dmg output of a hit is proportional to the max damage of the weapon
and the special you use. I verified this as a CL 80 char vs CL 80 gurks.

From the swordsman special post and my experiments you see that different
special just have the same old damage multiplier. (1.0 for melee hit, 1.7 for
Impr. Head hit in my case)

let:
p1 = proportion of time I use Head Hit
p2 = proportion of time I use Melee Hit

rps = Rate of action regen per second

spd = Modded Speed of weapon (from examine window of the charater in question)

maxdmg = Max damage of weapon

dmgmod = 0.45 (just I ball park number I chose which when multiplied by
maxdg gives base hit power)

sac = sac of the weapon

hit1cost = action cost of doing head hit per sac (0.15000)
hit1dmgmult = special dmg multiplier of head hit (1.7)

hit2cost = action cost of doing melee hit per sac (0.02)
hit2dmgmult = special dmg multiplier of melee hit (1.0)

So you get the following:

you pick p1 and p2 to maximise your action usage given your action recharge:

This is the action cost eqn:

rps = (p1 * hit1cost + p2 * hit2cost) * sac / spd

The amount of damage you do per second (dmg) is:

dmg = maxdmg * dmgmod * (p1 * hit1dmgmult + p2 * hit2dmgmult) / spd

and

p1 + p2 = 1 (ie. you do either head hit or melee hit always)

solving:

p2 = (hit1cost * sac - rps * spd ) / ((hit1cost - hit2cost) * sac)

p1 = 1 - p2

dmg = maxdmg * dmgmod * (p1 * hit1dmgmult + p2 * hit2dmgmult) / spd

Now this shows that

dmg is approximatly proportinal to

maxdmg / (sac * spd)

( Use this to rate your weapons!!! )

as long as:

1) sac is not so low enough that you can spam your most damage causing special
all the time ( sac >> rps * spd / hit1cost )

2) sac is not high enough that you cannot do melee hit constantly
( sac <= spd * rps / hit2cost )

3) spd not so high that you can do best specials all the time
( spd << hit1cost * sac / rps )

4) spd not so low that doing "Melee Hit" will drain action faster than you can
regenerate it ( spd >= hit2cost * sac / rps )

( note: 1 and 3 and 2 and 4 are just same equations just rephased )

( not in SWG land spd is actually weapon delay, but we all know this! )

Have more info and explanations, but not enough time atm.

Pretty interesting. I need to think about this some more, but your rating formula looks decent to me--certainly better than est. base DPS. In any case, it will still depend on how much a character relies on special moves.

Notably absent from your calculations is accuracy.

Yer, a weapon's accuracty does change, but only for different types.
A given weapon type has a fixed accuracy? (I think)

All it is is a factor to be multiplied after like the dmgmod number which I used above.
So that should not skew the results.

Simplified formula a bit to:

sac_coef = hit2dmgmult * hit1cost / (hit1cost - hit2cost)
- hit1dmgmult * hit2cost / (hit1cost - hit2cost);

spd_coef = rps * ( hit1dmgmult - hit2dmgmult ) / (hit1cost - hit2cost);

dmg = maxdmg * dmgmod * ( sac_coef * sac + spd_coef * spd ) / (spd * sac);

Variables the same as above.

( doing all this in octave - a matlab clone! - a free mathematics package )

So you see it is maxdmg / (spd * sac) plus a modfier depending on your specials

(
for my improved head hit and melee hit it is
sac_coef = 0.89231
spd_coef = 14.808
so damage output is more sensitive to sac than spd
)

This is all dependant on how you fight, but I do think (I have some evidence in a octave script) this is the highest damage output method I can think of.

(neglecting state attacks which I see as not changing the result, you just have to apply a state attack special every so often)

Do you have a link to octave? (too lazy to search for it right now).

Anyway, I'm not sure since your formula might overrate the importance of SAC. We might need to come up with some sort of normalization scheme to account for that. On the other hand, if you are just comparing weapons of the same class (pistol, rifle, etc) then normalization probably doesn't matter.

Iwill still need to read it a couple of times hehe not a total math guru. But the formula looks oke. That is to compare the same type of weapons you won't be able to compare a pistol vs a powerhammer. Also some people might go for the highest damage all the time. But my alt is a mbh/mp and 0430 carb i think. I also like to use snare and root and stun(slows down the regen rate) might even do some blinding to make them hit less and increase my accuracy with some crispic. It's a little like dps.

The dps formula uses the min/max damage weapon speed and your character modifiers. Offcourse we found out min damage is not used when you are in combat only max damage makes the difference. Maybe it's a bug maybe it's not. Either way it's broken and i don't think they will fix it anytime soon.

Now offcourse other specials will help in the amount of damage you do or the time it takes you to kill something. Bleeding,stun,dizzy(dunno?),armorbreak and that sort of specials.

Right now i'm working on a big test based on a dl44 xt pistol(easy and cheap to make) since most people are doing attacks every 2 seconds. I can do attacks every 1 second sometimes. At least that's what the combat window shows me and quickdraw is that attack. Now we don't really know how fast we attack yet since the combat window rounds the numbers up to the nearest second. It could either be 1.4 or 1.000001 or other attacks could be 1.500000 or 2.4 it might be a smaller range but we just don't know that for sure. I'm going to make my test and i'm going to try to figure out what weapon speed i need to attack every 2 seconds and 3 seconds i'm hoping i can figure out some more about the damage.

But basicly what i'm trying to see keep working on it. We need people like you to figure stuff like this out and eventually i think we will come up with a formula to compare most weapons equally.

Octave under download, Windows

The full script can be found at swg.m

( good read as has all equations in it )

In my last post I misread the equations

Graph

Shows how dmg varies with rps (regen per second), modifed weapon speed and sac. You see dmg is much more sensitive to spd than sac.

Yes states are important, but this is only really calculating max damage per second achievable for a given charater, with no states applied. That should be a good metric to rate a gun on (DPS has the same idea also)

Also I have hit a problem, the modifed weapon speed does not match hit frequency in combat log.
I assume melee hit an ranged shot do not have cooldowns.
(you can see this if you take speed buffs to a already fast weapon, you still hit faster so you are not cooldown limited)
So I need to find the equation which links modified weapon speed to hit rate in combat log.

Quite alot depends on attack speed, so it really needs to be figured out, may just boil down to trying lots of different weapons speeds hitting things a hundred times and taking the average time per hit.