Attack speed 2.0

[MaxQuest]

Ok, I understood what happens when you have two weapons equiped. When a single weapon is equiped you have 100% recovery. While using two weapons the game removes 50% recovery and you're left with 50% recovery for each hand.

Yeap. Have arrived to similar conclusion. If instead of [swaping base_recovery with dw_base_recovery] they have just made that [there is an inherent (and additive) 50% bonus when dw], majority of tests start to fall into place, without introducing the "double effect".

 

There are few small things, re-testing which wouldn't hurt through. For example:

- 1h recovery of 32, should be reduced to 16 frames, and not 15.

- dw + gloves (15%) + swiftaim (20%) + 2wpntln (20%), should result in 1.5 + 1.15*1.2*1.2 = 3.156. I.e. it should achieve no recovery with 15% malus armor; but there is actually 1-2 frames of recovery left.

But it's quite likely to be just rounding errors.

Edited March 29, 2016 by MaxQuest

[AndreaColombo]

So in absence of an attack speed penalty, a dual-wielder should reach 0 recovery with but a potion of DAOM. Do they?

[Kaylon]

So in absence of an attack speed penalty, a dual-wielder should reach 0 recovery with but a potion of DAOM. Do they?

Yes.

 

And with any 2 durgan enchanted weapons you can reach perma 0% recovery: 1.2 (style) x 1.15 (gauntlets) x 1.15 (durgan) = 1.59 while wearing durgan enchanted padded armor.

Edited March 29, 2016 by Kaylon

[Elric Galad]

So basically, with Durganized Plate, Dual Wield, Vulnerable attack, you will have 105 recovery with Dual wielding weapons.

That is exactly compensated by Durgan Weapon, and DAOM.

 

I guess Devil's too heavy armor is now too light...

 

 

So 1 hander with Speed mod are now suboptimal for dual wield.

[Kaylon]

So basically, with Durganized Plate, Dual Wield, Vulnerable attack, you will have 105 recovery with Dual wielding weapons.

That is exactly compensated by Durgan Weapon, and DAOM.

 

I guess Devil's too heavy armor is now too light...

 

 

So 1 hander with Speed mod are now suboptimal for dual wield.

If you want perma 0% recovery and more DR you'll need also speed on weapon. And I believe Devil can't use potions,food, drugs...

[Elric Galad]

Mmm more DR than a full plate ?

By the way my calculation was only about uber-optimization, so you can still use speed weapon, they will help for most fights.

 

Devil can drink potions from what I read. And she can also cast DAoM from twin stings.

[Kaylon]

Mmm more DR than a full plate ?

By the way my calculation was only about uber-optimization, so you can still use speed weapon, they will help for most fights.

 

Devil can drink potions from what I read. And she can also cast DAoM from twin stings.

I was talking about perma 0% recovery (ie no potions/buffs). Of course, if you have alacrity speed on weapons isn't required.

[hilfazer]

So in absence of an attack speed penalty, a dual-wielder should reach 0 recovery with but a potion of DAOM. Do they?

My spiritshifted druid indeed only needs DAOM to reach 0 recovery.

 

Hopefully crybabies won't discover it and start calling spiritshift OP.

[Elric Galad]

By the way, do anyone know what is the attack speed and recovery mechanic for full attack ?

[Boeroer]

 

So in absence of an attack speed penalty, a dual-wielder should reach 0 recovery with but a potion of DAOM. Do they?

My spiritshifted druid indeed only needs DAOM to reach 0 recovery.

 

Hopefully crybabies won't discover it and start calling spiritshift OP.

 

 

 

So in absence of an attack speed penalty, a dual-wielder should reach 0 recovery with but a potion of DAOM. Do they?

My spiritshifted druid indeed only needs DAOM to reach 0 recovery.

 

Hopefully crybabies won't discover it and start calling spiritshift OP.

 

With Two Weapon Style a Potion of Power or Outlander's Frenzy would be enough I guess?

 

Does this work: wear Sanguine Plate, walk into a fight, receive a crit and then spiritshift? Will Frenzy still be on? I think it will, but does anybody now for sure?

[Blunderboss]

Potion of Power and Outlanders Frenzy wont stack most likely , or does it ?

Edited March 30, 2016 by Blunderboss

[AndreaColombo]

They won't stack, but individually they should be enough if you're not using Vulnerable Attack. But why wouldn't you?

[Elric Galad]

 

 

So the 50% cut to recovery from when you dual-wield stacks additively with every other bonus?

Yes

Does it work the same with Bash shields ?

 

I tested that shield Bash is even faster than fast weapons (when used with one of them, the bash part of the cycle seems fasters) but I forgot to check how recovery works.

 

For me, it should be the same as Dual Wield except Dual Weapon Style does not work.

[MaxQuest]

Does it work the same with Bash shields ?

 

I tested that shield Bash is even faster than fast weapons (when used with one of them, the bash part of the cycle seems fasters) but I forgot to check how recovery works.

No, you don't get +50% additive bonus, for [1H + BashingShield] like you would for dual-wielding. Your main-hand's recovery remains the same.

It's just that bashing shield itself has the same attack_duration as a fast weapon, and a twice shorter recovery_duration.

Edited April 2, 2016 by MaxQuest

[Elric Galad]

Consistency, yeah !

[ancalimohtar]

 

Ok, I understood what happens when you have two weapons equiped. When a single weapon is equiped you have 100% recovery. While using two weapons the game removes 50% recovery and you're left with 50% recovery for each hand.

Yeap. Have arrived to similar conclusion. If instead of [swaping base_recovery with dw_base_recovery] they have just made that [there is an inherent (and additive) 50% bonus when dw], majority of tests start to fall into place, without introducing the "double effect".

 

There are few small things, re-testing which wouldn't hurt through. For example:

- 1h recovery of 32, should be reduced to 16 frames, and not 15.

- dw + gloves (15%) + swiftaim (20%) + 2wpntln (20%), should result in 1.5 + 1.15*1.2*1.2 = 3.156. I.e. it should achieve no recovery with 15% malus armor; but there is actually 1-2 frames of recovery left.

But it's quite likely to be just rounding errors.

 

 

Your second example in my spreadsheet creates 1.4 frames of recovery:

 

0.15 [armor] + 1-(1.15 [gauntlets] * 1.20 [swift aim]) + -0.20 [TWF talent] + -0.50 [inherent TWF] = -93% recovery time, which on a 40 recovery weapon would be 1.4 recovery frames.

 

 

By the way, what kind of one handed ranged weapons are you dual wielding? Are there any in the game? How else are you using Swift Aim with TWF?

Edited April 2, 2016 by ancalimohtar

[MaxQuest]

Your second example in my spreadsheet creates 1.4 frames of recovery:

 

0.15 [armor] + 1-(1.15 [gauntlets] * 1.20 [swift aim]) + -0.20 [TWF talent] + -0.50 [inherent TWF] = -93% recovery time, which on a 40 recovery weapon would be 1.4 recovery frames.

Thanks for point it out. Consider it proved that the bonus from TFW Talent stacks additively.

 

By the way, what kind of one handed ranged weapons are you dual wielding? Are there any in the game? How else are you using Swift Aim with TWF?

I dual-wielded... daggers! Swift aim actually works with melee weapons too.

 

 

 

 

 

 

But in case it wouldn't... I would probably give Long Pain a try: (px2_the_long_pain_weapon)

 

 

 

 

 

Edited April 2, 2016 by MaxQuest

[MaxQuest]

Good news guys,

I've managed to patch the assembly dll and log the exact values for attack and recovery durations.

Here they are (for some weapons, I had time to check so far):

 

 

Few interesting things:

- the game doesn't seem to use "frames" anywhere in the code. All durations are in seconds.

- there is a consistent relation between attack and recovery duration, namely recovery factor.

 

RecoveryFactor = 1.2

Single1HWeapRecovFactor = 0.5

 

- When you dual-wield, the default factor is used, and attack duration is divided by it in order to get the recovery duration (e.g. dw sabres: 1/1.2 = 0.833; dw daggers: 0.666/1.2 = 0.555).

- When you have only one weapon equipped, an additional division by 0.5 is performed (e.g. War Bow: 1.5/1.2/0.5 = 2.5)

 

And to answer one of my previous questions ("why does DEX doesn't affect reload duration uniformly, and why I was often getting results 0-12 frames off"): Because reloading animation is a looped one. The game periodically calls an Update() method and checks how much animation is there left to play. If it is less than hardcoded 0.5s than it sets the loop to false and plays one fragment more. Thus was emerging a possible 0-15 frames error in reloading duration.

 

 

ToDo:

- lookup for the inter-action delay.

- check if attack phase has "animation-rounding" mechanism somewhere, like reloading phase does.

Edited May 3, 2016 by MaxQuest

[Kaylon]

Then the base recovery is attack/0.6 and if you DW then the recovery is cut in half and becomes attack/1.2.

 

In terms of frames att/rec we have:

20/33 for fast melee weapons

30/50 for slow melee weapons and fast ranged weapons

45/75 for slow ranged weapons

and to that you have to add 7-8 frames per cycle from the inter-actions delays.

 

But we still need info about reloading which was always the hardest part to define.

Edited May 3, 2016 by Kaylon

[MaxQuest]

Then the base recovery is attack/0.6 and if you DW then the recovery is cut in half and becomes attack/1.2.

Yeap. It's easier to note it that way.

Or x5/3 for single weapon. x5/6 for dw.

 

 

 

Worth mentioning: in AttackData.cs there still can be found old, legacy values, e.g:

RecoveryFactor = 1

Single1HWeapRecovFactor = 2

But they are getting overridden by the values from global assets. (e.g. 1.2 and 0.5 respectively; and probably multiplication was substituted with division somewhere)

 

In terms of frames att/rec we have:

20/33 for fast melee weapons

30/50 for slow melee weapons and fast ranged weapons

45/75 for slow ranged weapons

To be more exact:

20/33 for fast melee weapons

30/50 for average melee, slow melee and fast ranged weapons

45/75 for average, slow, very slow and extremely slow ranged weapons

 

and to that you have to add 7-8 frames per cycle from the inter-actions delays.

Yeah. To stay in touch with empirical tests, it should be 6 frames for daggers and 7-8 for sabres/ranged weapons...

 

But we still need info about reloading

Found these values in the weapon asset files:

Arquebus: float ReloadTime = 6.800000 (or 204 frames)

Arbalest: float ReloadTime = 6.000000 (or 180 frames)

Blunderbuss: float ReloadTime = 5.000000 (or 150 frames)

Crossbow: todo

 

For reference, frapsing has yielded:

Arquebus: 189 frames

Arbalest: 167 frames

Blunderbuss: 150 frames

And 0-15 frames error stays in limit of the aforementioned 0.5s threshold.

Edited May 3, 2016 by MaxQuest

[Elric Galad]

Thank you for your work.

 

Do you think you could also hack pet and spiritshift attack speeds ?

 

Pets were said to have different BASE attack speeds (independantly from Dex).

 

I suppose forms attack should work like average dual wield, but I would be curious about cat form.

[MaxQuest]

Pets were said to have different BASE attack speeds (independantly from Dex).

Have tried a few. It seems that it was said wrong)

 

 

- Pet attacks are considered single-weapon

- Pet attacks suffer from a 40% armor penalty

- They do benefit from Dex

- Different companions have different Dex. Exact stats here.

 

I suppose forms attack should work like average dual wield, but I would be curious about cat form.

Indeed they are considered DW and cat/stelgaer forms attack the fastest.

Although boar form potentially might be also worthy of class builders' attention, due to it's damage over time.

 

 

P.S. More registered weapon data:

 

 

 

 

Edited May 3, 2016 by MaxQuest

[Elric Galad]

Once more, thank you for your testing.

 

And sorry, MaxQuest. I should have been more specific in my statements.

 

I had read that Lion and Bear (and only them) had about 33% slower attack speed, while boar, antelope, wolf, Itumak and stag had similar attack speed.

 

There seemed to be exactly 2 groups. Given that most pets (except wolf and Itumaak) had the same damages, lions and bears were a bit gimped. Slight differences could be explained by Dex. It was an old topic. I could try to find its trace. You have unfortunately only picked pets from the same group. So your testing is pretty much consistent with what I had in mind (except I had no proof that all pets from same group had exactly the same attack speed).

 

 

Aren't Itumaak stats supposed to be exactly the same as wolf ? I was surprised by the 15 dex from your testing.

Edited May 3, 2016 by Elric Galad

[MaxQuest]

You are welcome

 

I had read that Lion and Bear (and only them) had about 33% slower attack speed, while boar, antelope, wolf, Itumak and stag had similar attack speed.

Huh, interesting... Have checked and can confirm that:

 

[Elric Galad]

You are welcome

 

 

I had read that Lion and Bear (and only them) had about 33% slower attack speed, while boar, antelope, wolf, Itumak and stag had similar attack speed.

Huh, interesting... Have checked and can confirm that:

 

Nice testing ! But the conclusion is really sad as it confirm my thought that Bear and Lion are gimped.