Weaponsmith Archive
Thread: DPS is worthless. Post-CU, only DPA matters
Page 1 of 4
Curby
Fri Jun 10, 2005 9:00 pm
#1
First point: DPS is worthless
Evidence: Consider how DPS is calculated. It takes the average of min and max dmg, and divides the result by the speed of the weapon.
Screenshot of example weapons: http://www.speakeasy.org/~curby/swg/2hammers.gif (not really a link. grr)
(480+1101)/2/2.79 = 283 Base DPS
(399+920)/2/2.85 = 231 Base DPS
So, DPS is calculated using weapon speed. Now consider weapon speed and cooldowns. There is a generic combat-action cooldown that affects all actions after any action is performed. Each action also has a group/specific cooldown so certain powerful attacks cannot be spammed continuously. If you spam a single powerful attack, you will not be able to spam it at your maximum speed because the specific cooldown is almost always longer than the generic cooldown. If you alternate between several attacks, you will be able to attack at maximum speed. Here I am using two different hammers (not the hammers used above) with 10 attacks, alternating between critical and head hits:
2.79s weapon, 10 attacks takes 32s
2.46s weapon, 10 attacks takes 32s
It takes 32 seconds to perform 10 attacks with weapons that the untrained eye would claim have significant speed differences (13% in fact). If the speed of the weapon really did matter, we should see a 13% or 4 second difference between the two. More generally and qualitatively, tests have shown that weapon speed doesn't actually hinder you if you alternate attacks.
So consider the DPS calculation again. All things being equal, higher damage is always good. But we have shown that all things being equal, a faster weapon doesn't make you attack more often if you alternate specials. Thus, weapon speed is useless, and the only part of the DPS equation that we really care about is damage. We can then disregard DPS completely and focus on damage.
Second Point: DPA is everything
Consider a fight in which the quality of your weapon really matters. What manner of fight is this? Your opponent will have CL roughly equal to or greater than yours. This fight will last several minutes, in which your action is bound to drop if you want to maximize the number of specials you land. In fact, in most high-end fights you will be able to drop your action faster than your opponent can drop your health. Obviously SAC is important. The lighter the weapon, the longer you can swing it continuously without emptying your action pool.
Think of action as your potential to deal damage. When you use a special attack, you use up some of your potential. Take a heavy, high damage weapon and a light, low damage weapon. The heavy weap will allow you to do fewer hits before your action runs out, but each hit does more damage. But which weapon maximizes your potential?How much damage can you deal per unit action? In other words, how much does SAC matter with respect todamage? I'll let the numbers do the talking:
CL80 vs 54 PVP target, damage is total before armor reduction
#1 399-920 84SAC (CU eto)
11-12% 856 dmg 71.3 DPA (damage per action) 1.2063 dmg factor
#2 384-894 86 SAC (CU mapiasal)
12% 830dmg 69.2 DPA 1.1684 dmg factor
#3 480-1101 104 SAC (pre-CU stettin)
15% 1026 68.4 DPA 1.1554 dmg factor
#4 417-952 94 SAC (CU eto)
13-14% 888 63 DPA 1.0714 dmg factor
#5 394-967 104SAC (pre-CU mapiasal)
14-15% 888 59.2 DPA 1 dmg factor
CL 80 vs 80 PVP, head hit, dmg is before armor reduction
#1 399-920 84SAC (CU eto)
838 dmg 12% 69.8 DPA 1.2068 dmg factor
#4 417-952 94SAC (CU eto)
870 dmg 13-14% 62.1 DPA 1.0739 dmg factor
#5 394-967 104SAC (pre-CU mapiasal)
868 dmg 15% 57.9 DPA 1 dmg factor
What does this all mean? We perform two sets of tests. First tests against a CL54, and then tests against a CL80. Both times we are performing head hits and recording dmg before armor reduction (to reduce the variables involved). Each test is performed several times to see the range of action used by the weapon being tested.
First line in each test is the weapon used. Weapon number, damage range, SAC value, and era of creation (pre/post CU) and smith. Second line lists the damage, action used, DPA (damage per percentage point of action), and damage factor. For DPA, the worst case scenario (most action use detected in any trial) is used.
What is damage factor? It is a comparison of different weapons against a baseline weapon (map's 104SAC weap in this case). It is significant in that we can see different weapons have very similar relative performance regardless of target and target CL, whereas dpa changes depending on target. In other words, regardless of target, eto's 84SAC weap consistently lets you do more damage as you use up action. Damage per attack is less, but damage over time is greater because you won't be action-gated as quickly or as often.
Conclusions: [Edited based on comments below]
Evidence: Consider how DPS is calculated. It takes the average of min and max dmg, and divides the result by the speed of the weapon.
Screenshot of example weapons: http://www.speakeasy.org/~curby/swg/2hammers.gif (not really a link. grr)
(480+1101)/2/2.79 = 283 Base DPS
(399+920)/2/2.85 = 231 Base DPS
So, DPS is calculated using weapon speed. Now consider weapon speed and cooldowns. There is a generic combat-action cooldown that affects all actions after any action is performed. Each action also has a group/specific cooldown so certain powerful attacks cannot be spammed continuously. If you spam a single powerful attack, you will not be able to spam it at your maximum speed because the specific cooldown is almost always longer than the generic cooldown. If you alternate between several attacks, you will be able to attack at maximum speed. Here I am using two different hammers (not the hammers used above) with 10 attacks, alternating between critical and head hits:
2.79s weapon, 10 attacks takes 32s
2.46s weapon, 10 attacks takes 32s
It takes 32 seconds to perform 10 attacks with weapons that the untrained eye would claim have significant speed differences (13% in fact). If the speed of the weapon really did matter, we should see a 13% or 4 second difference between the two. More generally and qualitatively, tests have shown that weapon speed doesn't actually hinder you if you alternate attacks.
So consider the DPS calculation again. All things being equal, higher damage is always good. But we have shown that all things being equal, a faster weapon doesn't make you attack more often if you alternate specials. Thus, weapon speed is useless, and the only part of the DPS equation that we really care about is damage. We can then disregard DPS completely and focus on damage.
Second Point: DPA is everything
Consider a fight in which the quality of your weapon really matters. What manner of fight is this? Your opponent will have CL roughly equal to or greater than yours. This fight will last several minutes, in which your action is bound to drop if you want to maximize the number of specials you land. In fact, in most high-end fights you will be able to drop your action faster than your opponent can drop your health. Obviously SAC is important. The lighter the weapon, the longer you can swing it continuously without emptying your action pool.
Think of action as your potential to deal damage. When you use a special attack, you use up some of your potential. Take a heavy, high damage weapon and a light, low damage weapon. The heavy weap will allow you to do fewer hits before your action runs out, but each hit does more damage. But which weapon maximizes your potential?How much damage can you deal per unit action? In other words, how much does SAC matter with respect todamage? I'll let the numbers do the talking:
CL80 vs 54 PVP target, damage is total before armor reduction
#1 399-920 84SAC (CU eto)
11-12% 856 dmg 71.3 DPA (damage per action) 1.2063 dmg factor
#2 384-894 86 SAC (CU mapiasal)
12% 830dmg 69.2 DPA 1.1684 dmg factor
#3 480-1101 104 SAC (pre-CU stettin)
15% 1026 68.4 DPA 1.1554 dmg factor
#4 417-952 94 SAC (CU eto)
13-14% 888 63 DPA 1.0714 dmg factor
#5 394-967 104SAC (pre-CU mapiasal)
14-15% 888 59.2 DPA 1 dmg factor
CL 80 vs 80 PVP, head hit, dmg is before armor reduction
#1 399-920 84SAC (CU eto)
838 dmg 12% 69.8 DPA 1.2068 dmg factor
#4 417-952 94SAC (CU eto)
870 dmg 13-14% 62.1 DPA 1.0739 dmg factor
#5 394-967 104SAC (pre-CU mapiasal)
868 dmg 15% 57.9 DPA 1 dmg factor
What does this all mean? We perform two sets of tests. First tests against a CL54, and then tests against a CL80. Both times we are performing head hits and recording dmg before armor reduction (to reduce the variables involved). Each test is performed several times to see the range of action used by the weapon being tested.
First line in each test is the weapon used. Weapon number, damage range, SAC value, and era of creation (pre/post CU) and smith. Second line lists the damage, action used, DPA (damage per percentage point of action), and damage factor. For DPA, the worst case scenario (most action use detected in any trial) is used.
What is damage factor? It is a comparison of different weapons against a baseline weapon (map's 104SAC weap in this case). It is significant in that we can see different weapons have very similar relative performance regardless of target and target CL, whereas dpa changes depending on target. In other words, regardless of target, eto's 84SAC weap consistently lets you do more damage as you use up action. Damage per attack is less, but damage over time is greater because you won't be action-gated as quickly or as often.
Conclusions: [Edited based on comments below]
Look to maximize DPAin any weapon you buy. Usually, and especially if you do not use powerups, this means first reducing SAC and then increasing damage as a secondary consideration. If you do use powerups, the powerup will affect your weapon buying decision. Some particular combination of a -SAC pup on a high damage, high SAC weapon might be the most effective. Alternately, a+dmg pup on a low damage, low SAC weapon might be more effective. This depends on the availability and relative quality of pups and weapons on your server.
Finally, if you see that you do not drop your Action bar even when spamming your most effective specials in an extended fight, you might consider a weapon/slice/pup/buff combo that is biasedmore to damage output than low Action use.
Thanks to Mapiasal of TUA and Eto of TCO on Chilastra for being both the weaponsmiths providing these weapons and the punching bags the weapons are used on.
Thanks to Mapiasal of TUA and Eto of TCO on Chilastra for being both the weaponsmiths providing these weapons and the punching bags the weapons are used on.
Message Edited by Curby on 06-10-2005 10:03 PM
Message Edited by Curby on 06-10-2005 10:07 PM
Message Edited by Curby on 06-13-2005 10:49 PM
Curby
Fri Jun 10, 2005 9:02 pm
#2
I just made a post on the weaponsmith forum discussing DPS and SAC. My tests use a series of Power Hammers, and I thought I'd post a blurb here as Swordsmen may be interested:
http://forums.station.sony.com/swg/board/message?board.id=weaponsmith&message.id=86448
Message Edited by Curby on 06-10-2005 10:08 PM
Trigg_Ichosw
Sat Jun 11, 2005 4:44 am
#4
Very good post. I did my own tests which i posted (http://forums.station.sony.com/swg/board/message?board.id=weaponsmith&message.id=85460), and your tests back up what i was trying to get across.
PaladinX333
Sat Jun 11, 2005 7:06 am
#5
I agree with the conclusions in this post. Having said that, DPS is a great marketing tool and I still sell weapons because they have a high DPS number. I can and do make low SAC weapons on request. For the masses however, DPS is king. I give the people what they want since it affects the bottom line directly.
Rico4
Sat Jun 11, 2005 7:19 am
#6
good post that should clear things up for good.
Alot of 'Smiths have been saying that sac is most important for ages-myself included, yet i still get people coming upto me saying i want a weapon with 330+dps. so i explain about sac and they say yeah ok that sounds better actually.
The way i see it is high dps is good for certain templates where you might have, sayranger,-which doesnt really give much speed so you can notice a difference with faster weapons, and maybe good for grinding against lowish level mobs.
you wanna take down something that hits hard and fast but only has 10k health then yes you can use a high damage, high dps, high sac weapon, but anything form say 17-50k health low Sac is essential! to survive.
Id say if your in agroup of 4 or less then highsac is a no go, bigger groups you will get away with a high sac weapon as you can 'rest' while the rest of group continues.
Anyway thats my thouhgts on this matter, now ill stop blabbering on 
Again, excellent original post 
Message Edited by Rico4 on 06-11-2005 03:21 PM
laodamas
Sat Jun 11, 2005 10:43 am
#7
Ive been saying this for weeks now, but people dont seem to get that against an opponet each special will do the same damage each time. This damage is not based on dps, its based off of max weapon damage. Speed shown on the weapon means very little compared to the speed you hit for with all things factored in. Eventually players will learn, but till they do feed them whatever they are willing to buy. I look at it this way, first time they buya high dps gun, realize they run out of action to fast. Second time they buy a weapon with a little better action cost, but still more interested in dps. Finally they learn and buy a weapon with low action cost and high damage, not really caring about how speed effects dps(because slicing and mods will lower speed anyway). Result I sell more weapons and make more credits.
D4rthReven
Sat Jun 11, 2005 1:47 pm
#8
with swordsmen this has been obvious since first day of cu hehe if you didnt notice that you did more damage with a weapon that had a reasonable action cost your blind 
WarmJast
Sat Jun 11, 2005 2:15 pm
#9
Very nice post Curby. I'm going to go bug a weaponsmith to make me a low SAC ph now and see how it turns out for me.
SlaughterMonster
Sat Jun 11, 2005 2:34 pm
#10
wow i'm glad people are finally starting to realize this.....unfortunately i posted this info in an auction i was having for some really uber pistols/carbs/rifles i had made and peolpe on my server had a hissyfit calling me a bogus BS WS.....so i took the weapons off and called everyone that hasn't bought custom orders from me an idiot basically lol....that didn't go very well
It is our job as WS's to pass along this info to our customers and make well rounded weapons and not having to make 2 types of weapons....elite and idiot weapons.
Curby
Sat Jun 11, 2005 8:48 pm
#13
laodamas wrote:
This damage is not based on dps, its based off of max weapon damage.
Not quite. Both min and max dmg matter.
Example:
CL80 vs 80 PVP, improved head hit
331-895 hammer does 775 dmg before armor reduction
384-894 hammer does 802 dmgbefore armor reduction
Average dmg of first hammer is 613
Average dmg of second hammer is 639
Difference of average dmg is 26
Difference of actual pre-armor dmg is 25
Page 1 of 4