Can normal attack cancel skill casting ?

freygin

Does anyone playing cleric or other class with skills that have long channel time notice that sometimes we can get cancelled when casting a skill ? I think it's normal attack that has that cancelling effect to us casters. (not barb's skill with cancel effect)

Aeliah - Dreamweaver

First, PLEASE don't get into the nasty habit of using 'channeling' and 'casting' interchangeably. Channeling is the initial phase of a skill, and interrupts can happen here. Casting is the second phase of a skill, and it can not be interrupted (in fact, if you are stunned or killed during casting, your attack still completes, which can lead to the phenomenon where you and your opponent die at the same time).

Next:

Heavy armor classes are relatively immune to having the channeling for their skills cancelled.

Light armor classes sometimes have skill channeling cancelled (by any attack).

Arcane armor classes often have skill channeling cancelled (by any attack).

There's an interesting question I myself would be interested in knowing the answer to:

Is the chance to have the channeling of a skill cancelled directly proportional to the length of the channeling, or does that chance vary by skill independently of the length of channeling?

Let me explain why I'm not sure either way. I've observed that, in Phoenix Valley with upwards of 50 mobs hitting me at once, it is fairly difficult to get a stream of rejuvenation (2.5s channeling) off without having its channeling cancelled. However, ironheart (1s channeling) often goes off without getting cancelled, which is odd---if the chance to have channeling cancelled was directly proportional to length of channeling, we'd expect 2.5x more cancels for stream of rejuvenation, but the gap feels much larger. Without having done more testing, it feels like stream of rejuvenation has at least a 5x higher chance of having its channeling cancelled.

Then take Tempest. Tempest has a 4 second channeling. In theory, if the chance to have channeling cancelled by an attack was proportional to channeling length, then it would be slightly less than twice as often as a stream of rejuvenation gets cancelled. In reality, however, firing off a tempest with 50 mobs hitting you is nearly impossible. It almost *always* gets cancelled. Isn't that odd? For this reason I feel that perhaps the chance to have a skill cancelled is specific to each skill.

If this is the case, then probably it isn't 'class differences' so much as 'skill differences'. In my theory, each skill has a % chance, call it X, to have its channeling cancelled when a hit lands on you during the channeling.

At the lower end, we'd have X = 0 for a barbarians's invoke. Unless somebody can tell me otherwise, I have never, ever seen invoke channeling cancelled by an auto attack/attack that doesn't specifically interrupt channeling or stun/sleep. Even when 100 mobs are hitting my barb, invoke goes off without fail.

On the other hand, a cleric's Bubble of Life (bb) seems to have a very high rate of having its channel cancelled. If I have 20 mobs on me, it can take quite a few tries before bb finishes channeling, due to frequent interrupts from normal mob attacks. Maybe the X factor is... (guessing)... 5%?

So yeah, if I'm right, there should be observable differences between skills for how often their channeling gets interrupted. I should do some testing with my sin and cleric... hmm... has anybody else done testing like this before?

Sakubatou - Sanctuary

Aeliah - Dreamweaver wrote: »

it is fairly difficult to get a stream of rejuvenation (2.5s channeling) off without having its channeling cancelled. However, ironheart (1s channeling) often goes off without getting cancelled, which is odd---if the chance to have channeling cancelled was directly proportional to length of channeling, we'd expect 2.5x more cancels for stream of rejuvenation, but the gap feels much larger. Without having done more testing, it feels like stream of rejuvenation has at least a 5x higher chance of having its channeling cancelled.

I'd guess channeling has a vulnerability point that isn't the entire length of the channeling (but I'm already rethinking this). You observed with your thread on when a stun lands about a .2 second discretion and attributed this to ping and I agree. The time when a skill is told to the start by you, the time the server approves it, and the time your computer kicks it into motion. I'd guess at the start of a skill while this happens and at the end of channeling as it kicks over to casting the skill is in an transition process where it can't be interrupted. In other words it's "starting to channel" but not really channeling yet because of ping. If you take away some of these processes but leave the middle chunk it makes a big difference on quicker skills.

For instance, lets say IH has .2 seconds at the beginning and the end where the game is saying it's going to be channeling or going to be casting and so isn't vulnerable to the channeling phase. This reduces it from 1 seconds to .6 Lets say SoR experiences the same thing, but it only gets reduced to 2.1 seconds. Now we have a 3.5x difference.

What makes me already rethink this concept that the transition processes of starting channeling or starting casting causing ping lag is "instant" skills have a .1 second channel and I've had them interrupted.

DarkSkiesx - Dreamweaver

A simpler explanation could be that perhaps the game registers hits on a interval basis. I.e. although there are 50 mobs chasing you the 50 hits register in say 0.5second interval or even 0.25 second clumps.

So under this theory you would have a chance of being interrupted every interval regardless of how many mobs are hitting you, and this may also explain why two hits registering at the same time (or interval) can charm bypass when you would expect your charm to tick instead after the first and then register the second.

Aeliah - Dreamweaver

Hmmm... by that logic we might expect the smallest interval to be 0.2s, because 5aps generates the interval time of 0.2. However, at 5aps on my sin I regularly bypass people's charms with dps. Lag? Design? Perhaps the interval is closer to 0.25 or 0.3s after all. Interesting...

I've had triple spark interrupted before too, its extremely rare, but definitely possible. Very odd indeed.