Dismissing the weapon speed gossip

[Alash]

Introduction
I can't even begin to comprehend the amount of times I've discussed weapon speed in LOTRO. Online, In private forums, and here on the public forums. Most agree with the basic concepts, that faster weapons means more auto-attacks, and slower weapons means higher damage through skill-based attacks. And then there's the details about power consumption, cooldown-timer cycles, and the very loose concept of certain weapon speeds giving the right 'feel'. What the discussions never brought, was conclusive arguments. Yes, these factors play a part, but how big a part? How does the higher skill-based attacks weigh against the increased amount of auto-attacks? Those are difficult questions to answer, but none the less that's exactly what I intend to do in this article.

Setup and particulars
An important condition for me to embark om this project has been that the test could be applied to actual gameplay. I wanted high end gear to go with a high-end weapon setup with the best legacies; legacies that needed to be identical on the weapons I wanted to test. Thanks to some luck and a fair bit of persistance I now have such weapons; an axe, and a sword. Coupled with the best crafting can offer, these are the two setups:





Other gear is as shown in the Damage Build in the Champ Level 60 Gear thread - as are traits. The test was accomplished with CStats Beta 5, pulling Ghâsh-hai Rampagers in Library of Steel, and against the first boss in Grand Stairs. Focus was on showing three aspects of combat: one for single-target damage, one for AOE, and one for endurance over the course of a longer fight.

The skill sequences used are as follows (BS = Brutal Strikes, SS = Swift Strike):

• Single target
  Axes: Wild Attack > Flurry > Wild Attack > SS > BS > Merciful Strike > Wild Attack > BS > Wild Attack
  Swords: Wild Attack > Flurry > Wild Attack > SS > BS > Wild Attack > Merciful Strike > BS > Wild Attack > SS > BS

• Two targets
  Axes: Bladewall > Rend > SS > Bladewall > Flurry > Blade-storm > Bladewall > BS > Merciful Strike
  Swords: Bladewall > Rend > SS > Bladewall > Flurry > SS > Blade-storm > Bladewall > SS > BS > Merciful Strike

• Arch Nemesis
  Axes: Flurry once every 30 sec, Ferocious Strikes once every 30 sec, Battle-Frenzy once every 60 sec, and a blend of Wild Attack, SS, BS, Feral Strikes and Merciful Strike
  Swords: Flurry once every 30 sec, Ferocious Strikes once every 30 sec, Battle-Frenzy once every 60 sec, and a blend of Wild Attack, BS, Feral Strikes and Merciful Strike

The first four tests (those in Library of steel) were done by killing 23 orcs, and removing the three that deviated most from the average (for some reason the DPS reports display my DPS in a factor 1000 billion. To counter this, I replaced the reported DPS figures with the real ones, rounded up to the closest full number).

The statistics





The single target data shows a clear distinction between the two setups. Even though the 20 sword entries applied on average almost 2% more damage, they delivered only 82,7% of the DPS axes did.





Again, the pattern is distinctive. Even though axes did on average 0,6% more damage (small inaccuracies in the timing of ending encounters within CStats), the swords fall far behind and offer only 82,2% of the axes' DPS.





And finally, we have the endurance test. As can be seen from the total damage the boss managed to heal in the first test (disarmed and used fear simultaneously, leaving pots on CD), making the encounters last almost exactly the same time. This is good because it enables us to compare the number of attacks.

If we look at the overall pattern, it is still persistent: swords could maintain only 85,4% of the DPS axes provided. Apart from being a rather clear advantage for axes, a fair few points can also be derived from the distribution between skills, and more importantly their share of the total output. For instance it may surprise some that auto-attacks for the much slower axes contributed almost 5% more to the total damage than they did for swords. Fewer attacks, but more auto-attacks. That sounds strange - or does it? With longer cooldown between each skill use, there is room for more auto-attacks to land, and since they work on a seperate cooldown timer than skills, there is - in this case - nothing gained from having faster weapons. With the swords, it seems the succession of skills are so rapid that they allow for less auto-attacks!

Another figure worthy of mention is the average damage. When looking at global weapon speed, swords are 18% faster. The average damage is 17% lower on swords. So despite the various static damage modifiers on skills, it would seem that there's a rather tight connection between weapon speed and average damage.

We can also try and look at the connection between attack speed and number of attacks. Swords are 18% faster, but only landed 6% more hits total. And at the same time, axes landed 22% more auto-attacks. A rather essential point to conclude from this, then, is that faster weapons allows for a more rapid succession of skills, and are thereby more prone to 'block' auto-attacks from landing in between these skills. Ironically, a commonly perceived advantage of faster weapons is actually a weakness.

And lastly, we need to consider the power consumption. I killed the arch nemesis five times before making the actual tests just so I could get used to the optimal skill sequences and learn the timing, but what I also figured out during this was that while axes could go flat out 75% through the fight, swords went dry of power only 50% through. Because of this, I had to severely reduce the amount of Swift Strikes with swords, which obviously made its DPS suffer. In this case, the power consumption was a third smaller with axes than with swords.

Errors and accuracy
Some things needs to be said about the validity and accuracy of the data. Of variables, we have the damage type of the off-hand sword, which orcs aren't weak to. The effect of this has previously been tested by others, and shown to be around 5%. Since the off-hand weapon provides a relative small fraction of the total DPS, the effect of this weapon type change is assumed to be marginal, at around 1-2%. Another inaccuracy, as mentioned earlier, is the variance in total damage against orcs in Library of steel, which is somewhere between 0.6% and 2%. However this work both ways, and therefore don't favor the sword setup. And finally we have what's probably the biggest source of error: the skill sequence. Due to the differences in weapon speed it was not possible to maintain the exact same sequence or the exact same distribution of skills; it had to be adjusted to work optimally with both of the setups. I don't consider that much of an issue though, as the exact same skills were used, and the sequences didn't differ much (especially not in Library of Steel). But the biggest inaccuracy is in the scarcity of data.

Slower is better
As can be seen from the above spreadsheets, there is a both consistent and noticeable difference in performance between a 2.50 speed setup and a 2.05 speed setup, in favor of the slower. Against single targets, swords offered 82,7% of the DPS axes did. Against two targets at a time, that relationship was 82,2%, and against the arch nemesis it was 85,4%. Against single targets, multiple targets, and over time, axes performed on average 19,8% better. That is a rather large difference for weapon setups that have previously been considered somewhat equal.

The reason for this, as we've already discussed earlier, is twofold

1. The more rapid succession of skills that comes with faster weapons prevent auto-attacks from landing, and thereby don't receive the necessary compensation for smaller skill-based hits, which makes them fall behind on overall performance.
2. Conserving power becomes a factor approximately two minutes out with swords, and three minutes out with axes (when going flat out). This naturally makes faster setups fall behind on longer fights, and more so for each minute.

Conclusion
We have learned in this article that there's more to weapon speed than one might think. We have shown, that slower weapons are better on short-term performance as well as long-term, and better by such a margin that it shouldn't be ignored. Those swearing by the look and feel of swords may well continue to do so, but they should know that they're doing it at the cost of a considerable amount of performance. And of course this isn't just about swords and axes - it's about showing that as a general rule, slower means better.

[rumple9]

Admirable effort Alash but the data is based upon a certain sequence of attacks and an assumption of the use of auto attacks.

The data is sterile and ignores the the fact that in real time fights most champs are going to open with rend (even against single targets) or most likely improved rend. This cannot be ignored.The sample is too small to be of any use. It also discounts the fact that with faster weapons a champ is likely to crit more (or better still dev crit ) during the course of the fight.

You also ignore the fact that in set up 1 you are using 2 beleriand weapons, but in set up 2 the o/f sword has ancient dwarf.

Further the dps figures on the weapon headings only relate to auto attacks but the damage range on set up 1 is far superior:


set up one
Leg axe 85-116 o/f peerless axe 79-107


set up two
Leg sword 67-96 o/f peerless sword 64-91

So even ignoring weapon speeds the average damage on set up one is substantially higher than set up two, notwithstanding the fact that set up two is hindered by ancient dwarf.

Further most champs are likely to be using feral strikes than brutal strikes, which I believe has already been proved to be superior. The mere use of BS corrupts this data because it has a high miss chance, which means there is a large element of luck in your testing.

To do this sort of study you need to use like for like weapons in terms of damage range and damage type (which I doubt is possible).

The mere fact that the substantially inferior set up in test 2, produced only 52.7 less total damage (8079.2 - 8026.5) against multiple targets, BUT 79.5 MORE damage against single targets (3981.3 - 3901.8) would actually lend me to believe that faster weapons are in fact better.

However I doubt The Watcher is going to be worried by either set up.


It would be interesting to see what the results would be with a larger sample and a peerless thain's axe in each set up in the off hand which I think is hindering the results here.

Anyway thank you for the post (and effort) which is likely to produce a lively debate.

[Duvaineth]

hmmm your holy quest still goes on

looking at it dmg wise the difference isnt huge. And you are also using this data based on solo mobs.

if you can and do the sword lovers a favour redo this test on a 100k+ mob
and then on a lvl 62-63 mob. And you will see swords shining there.

And while i also agree with rumble...i disagree on that the watcher isnt going to worry
Watcher should fear swords and so should any raid boss above our level. increase chance to hit is very important fact in end game dps

[nonsferatou]

Guys thats an efford ,maybe the database has flaws but at least is something ....20% difference is not few ,neither is 8...
To sword lovers: prove what you saying ,cause this is a kind of proof..you will kill solo mobs 80% of the time ...bosses are once in a week and there is still a doubt for what you saying ....
p.s.How about testing a axe /sword comination or axe/hammer ?

[Alash]

Few points to make.

1) If you look closely I did open every fight with Rend. If you have done the math you will realize that Rend doesn't pay off on less than two targets, and thus I only used it on two targets.

2) Talking sample size, the deviation is around 6% which I've accounted for in my conclusion, so you can't say that it's too small. Had I done more tests that number would shrink to maybe 4%, but I don't see the point of doing that. The fact of the matter is that the margin of error is substantially smaller than the difference in DPS.

3) Faster weapons will make you crit more, and they did crit more. However, slower weapons crit for more which makes it somewhat of a moot point.

4) As for damage range, I don't see what you're getting at. With a slower weapon of the same DPS, damage range naturally has to increase, which is the whole point of this test - whether that increase makes up for the loss in speed.

5) Yes, Thain's Sword only has Ancient-Dwarf, and that affects performance, but to my knowledge that effect is less than 5%, and talking on the off-hander here that's a lot less on the total result. Since the vast majority of your DPS comes from main hand, the effect is probably down to 1-2%. However, that should indeed be considered, so I'll add that to the article.

[Duvaineth]

well yeh, but grinding solo doesnt come attached with a performance score...at least for me xD

So i might aswell look good and have a nice feeling while doing it hehe.

But yes the data and Alash continuing to be able to run tests for us is great, i dont think nobody is against that

[Amphoras]

Originally Posted by rumple9

The mere fact that the substantially inferior set up in test 2, produced only 52.7 less total damage (8079.2 - 8026.5) against multiple targets, BUT 79.5 MORE damage against single targets (3981.3 - 3901.8) would actually lend me to believe that faster weapons are in fact better.

It would be expected that they do the same total damage as both times you are killing the same number of the same mobs. If the total damage was hugely different, therre would be something very wrong.

[Iorielin]

Originally Posted by rumple9

Further the dps figures on the weapon headings only relate to auto attacks but the damage range on set up 1 is far superior:

That was just the stupidest thing I ever read! No offence.
If they had the same damage range and the sword would still have much faster speed then the axe it would put the DPS of the weapon at 50 or something. And this is all about DPS! Same DPS on both weapons (on paper), diffrent speed and damage range. But do they really have the same DPS in reality?

Alash research shows that the DPS is far superior on axes. I always knew they were better, but I didnt think it was such a big differnece, so I've been wielding swords so far. The total damage done is irrelevent as his killing the same mobs. The axes had superior DPS meaing he killed the mob faster, he took fewer hits, he had more morale left when the fight was over, and on top of that he used less skills meaning he had more power aswell.

This really made me think about changing from a sword setup to a axe setup. It doesnt look as cool on a elf with axes, but atleast I preform better.

[Duvaineth]

Originally Posted by Iorielin

This really made me think about changing from a sword setup to a axe setup. It doesnt look as cool on a elf with axes, but atleast I preform better.

on solo mobs /short fight mobs :P

[Alash]

Originally Posted by Duvaineth

on solo mobs /short fight mobs :P

What reasoning would lead you to believe that this doesn't apply to elite mobs and long fights?