Attack Animation

[Wolfos]

The Question

I'm trying to figure out if Attack Animations play into attack speed, because it seems like they should to me. Everyone tells me that they scale with attack speed, but how do they scale? If they scaled equally, then there would be no advantage to having faster animations.

So, let's run some numbers and see what we find.

Initial Scenario:

Let me illustrate my question with an example. *Let's go get Kunkka's numbers:
BAT: 1.7
Attack Animation: 0.4/0.3

Now let's go get another hero for comparison. *Let's use Doom:
BAT: 1.7
Attack Animation: 0.5/0.7

Let's assume that each hero has an IAS of 100 to make any math easy. *Let's also assume that no animations get cancelled for purposes of discussing times.

The Discussion

Ok, so according to Mechanics, the time in seconds between attacks is given by [ BAT / ( 1 + IAS ) ], so this equates out to 0.85 seconds for an IAS of 100 with or heroes' BAT of 1.7. *This is quick and easy, and it makes sense.

When trying to figure out where Attack Animations play into this and how to calculate it, I saw this: *http://www.gosugamers.net/dota/thread/767978/start/20 (post #25 is the one in question).

According to this, I can calculate how long it takes for my damage to take effect by using the formula [ Damage Point / ( 1 + IAS) ]. *It also says that I can calculate how much time is used by the backswing by using [ Backswing / ( 1 + IAS ) ].

For our examples of Kunkka and Doom, this says that Attack Animation Times come out to:

Kunkka
Damage Point: 0.20 seconds
Backswing: 0.15 seconds
Total Time: 0.35 seconds

Doom
Damage Point: 0.25 seconds
Backswing: 0.35 seconds
Total Time: 0.60 seconds

Looking at this, this doesn't seem to add up at all. *We already calculated out that our Attack time should 0.85 seconds, and these numbers definitely don't add up to that. *Is the guy from gosugamers wrong on his formulas or did I misuse them somehow?

Assuming that the formulas are all correct, that would be that there is a measurable dead time where your hero is unable to attack. *For Kunkka, this is 0.50 seconds, while Doom has 0.25 seconds. *

Let's see what happens when we set their IAS to 300. *Recalculating their Attack Time for BAT, we come up with 0.425 seconds per attack.

Now, let's rerun the calculations for Damage Point and Backswing.

Kunkka
Damage Point: 0.100 seconds
Backswing: 0.075 seconds
Total Time: 0.175 seconds

Doom
Damage Point: 0.125 seconds
Backswing: 0.175 seconds
Total Time: 0.300 seconds

There is still measurable dead time. *Kunkka is standing at 0.250 wasted seconds, and Doom clocks in at 0.125 wasted seconds.

So, in our initial scenario, there is no effect on Attack Time by Attack Animation.

Let's examine some extreme examples. *There are precious few heroes with lower than 1.7 BAT, but lets examine the numbers from Anti-Mage and from Alchemist while under his level 3 Chemical Rage.

Second Scenario

Anti-Mage
BAT: 1.45
Attack Animation: 0.3 / 0.6

Alchemist
BAT: 1.15
Attack Animation: 0.35 / 0.65

Let's go ahead and jump to the fun stuff and assume 300 IAS. *Using the same formulas as before, let's calculate the Attack Time, then the individual Damage Point and Backswing times.

Anti-Mage
Attack Time: 0.3625 seconds
Damage Point: 0.075 seconds
Backswing: 0.15 seconds
Total Attack Time: 0.225 seconds
Wasted Time: 0.1375 seconds

Alchemist
Attack Time: 0.2875 seconds
Damage Point: 0.0875 seconds
Backswing: 0.1625 seconds
Total Attack Time: 0.25 seconds
Wasted Time: 0.0375 seconds

So, we can see that there is always some amount of wasted time. *This shows that Attack Animation does not currently affect Attack Time.

Just for shits and giggles, let's look at the version of Anti-Mage with 1.35 BAT.

OP Scenario

Attack Time: 0.3375 seconds
Damage Point: 0.075 seconds
Backswing: 0.15 seconds
Total Attack Time: 0.225 seconds
Wasted Time: 0.1125 seconds

Not even that broke the barrier. *We can now conclusively see that Attack Animation has never affected Attack Speed, but we can use these calculations to show how efficient heroes are with their attacks as well as better understand how quickly they hit. *However, that's a quest for another day. *

[Erios]

Going with no IAS to help you understand.

Hero with 0.5/0.5 animations. 1.7 BAT.

Hero is chasing enemy unit.
Hero is in range and starts attacking.
0.5 seconds later attack is launched. - Backswing starts.
0.5 seconds later backswing ends. - Hero resumes chasing, if enemy unit is in range hero stays there.
0.7 seconds later if hero is in range second attack starts.


Say now our hero gets his animation canceled at 0.3 seconds.

Hero is chasing enemy unit.
Hero is in range and starts attacking.
0.5 seconds later attack is launched. - Backswing starts.
0.3 seconds later backswing ends due to animation canceling. - Hero resumes chasing, if enemy unit is in range hero stays there.
0.9 seconds later if hero is in range second attack starts.


Now first scenario but with a hero who has 100% IAS.

Hero is chasing enemy unit.
Hero is in range and starts attacking.
0.25 seconds later attack is launched. - Backswing starts.
0.25 seconds later backswing ends. - Hero resumes chasing, if enemy unit is in range hero stays there.
0.35 seconds later if hero is in range second attack starts.

[Wolfos]

^That's useful, but not really the point of the question. I was wondering how close some of the heroes came to breaking through the BAT barrier.

[Damberna]

It doesn´t affect attack speed

[Erios]

Originally Posted by Wolfos

^That's useful, but not really the point of the question. I was wondering how close some of the heroes came to breaking through the BAT barrier.

Death Ward.

[Sarrink]

What's your question exactly?

[Wolfos]

The initial question is: Do Attack Animations affect the Attack Speed of a hero?

I exhaustively prove that it's a no. The post was more of an informative discussion and not an actual question.

[Erios]

They might do, Death Ward is an example.

[JaCKaSS_69]

Attack backswing doesn't get reduced by increasing your attackspeed. It simply gets canceled (as in overwritten) if the attack cooldown is lower and another attack is following.

Attack point does indeed get reduced by increasing your attackspeed. This only makes sense because if it wasn't reduced for any hero (in DotA that is) who would have an attack animation higher than 1/5*BAT his maximum AT would be his attack point.

Btw, AttackPoint+AttackBackswing != Attackcooldown. As in the attackpoint and backswing of heroes doesn't always add up to be equal to the cooldwon of their attack.

[Heraklinos]

Originally Posted by JaCKaSS_69

Attack backswing doesn't get reduced by increasing your attackspeed. It simply gets canceled (as in overwritten) if the attack cooldown is lower and another attack is following.

/o/

Originally Posted by EebstertheGreat

The attack point represents the time it takes from the beginning of an attack to the release of a projectile (in the case of most ranged attacks) or the damage (in the case of melee and instant attacks) with 0 IAS. The backswing represents the time a unit ordered to do a single attack will continue animating after the attack point before proceeding to the next order (unless it is canceled) at 0 IAS. Both of these values scale normally with IAS, just like BAT.

[1239]

Attack backswing does scale with increased/decreased attack speed.

However JaCKaSS_69 is right that it may get 'canceled' when another attack is following because attack backswing is in all cases limited by following formula to a range of

0 <= attack backswing <= 1/attacks per second - damage point


@OP: as attack backswing is capped anyway and damage point scales with IAS the animations (as you call them) do not interfere with the attack speed. Only when the damage point is greater than 1/attacks per second with the allready given example Death Ward.
However when we talk about complex scenarios like chasing or generally ordering other commands it comes to differences in damage output for different animations.


@JaCKaSS_69: I like your sig, though few will remember its origin I guess^^

[JaCKaSS_69]

@1239:

Hi there! o/

I thought I've seen it back in DA forums that attack backswing wasn't affected with IAS but instead canceled when the next attack started. I guess it was a long time since I've been consistently lurking and searching mechanics.

Also, @sig, I remember back in the day, when I first saw it, I wholeheartedly and totally loved it!