Wereform Attack Speed Formula

[Dagni]

Now I've got it. The one thing I'd completely forgotten about with Zeal/rollback animations, and wasn't brought up by anyone in this thread, was that the speed formula doesn't have the '- 1' at the end when figuring a rollback attack, like Zeal. I.e.

This: Frames = {256*(Base + 1)/[(100 + Speed Increase)/100*Animation Speed]}
Not: Frames = {256*(Base + 1)/[(100 + Speed Increase)/100*Animation Speed]} - 1

So after changing that, I found that using '6' for WereBase fit the data I compared with. So 6+1 = 7 frames_to_rollback. Since Werewolf animation_length = 13, that would normally be expected to be 13/2 = 6. Which means that either I don't understand Ruvanal's explanation of rollback theory as well as I thought, or that werewolves do their actual attack just later in the animation than is normal (Frame 7 instead of frame 6).

So here is what the full formula looks like for Fury, already including the appropriate numbers for being a wolf (not bear) and the frames number for Fury.

delay = [256*(WeaponBase + 1) / [(weapon ias+wsm+100)*256/100]]

Anim speed = [9 * 256 / delay]

Speed increase = [120 * item_ias / (120 + item ias)] + skill ias + wsm
if (Speed increase > 75): Speed increase = 75

Frames = {256*7/[(100 + Speed increase)/100*Anim speed]}

Edit: In comparison, only difference for Wolf's standard attack is the final part:

Frames = {256*13/[(100 + Speed increase)/100*Anim speed]} - 1

- Dagni

[Dagni]

EDIT #3: I've re-worked my calculator to take into account what's been discovered here.  I've been going through the tables looking for errors but I haven't found one so far!  If you'd like to check out the info easily, check out my calculator.   :)  Note that the Fury data points still read off of my charts and so it still has limited data points.  Again, great job, guys!

Wereform Calculator

Nice work on the calculator. I don't know if I never read edit #3 before, or if my brain just never bothered to compute it, but either way I was pleasantly surprised when I went to Wereform Central to pull up the Fury chart I was checking the formula against, and saw the changes.

If you, or someone, could test even just a few data points for Hunger with both the wolf and bear, that might allow me to come up with at least a working theory on what the formula for that is. Because right now I have absolutely no hard data to compare formula results with. My current guess for Werebear; based upon it being 'about a frame faster' at first, but maxing out as fast, not faster; is an animation length of 13 (from Ruvanal), so a WereBase of '12', and a value for 'prev frames' = 11. For Wolf, change to: WereBase = 9, 'prev frames' = 7. Those are just guesses.

I'd do some testing myself, except I don't HAVE a character with Hunger/Rabies! :(

- Dagni

[TheDragoon]

Sure, I could test some data points if you'd like, but give me a week to do it. :) If you want to download a test character, I have my tester online for download here. Sometime today, likely, I'll input the Fury formula and check it against existing charts just to be sure that it works. B)

[TheDragoon]

Awesome, I'm convinced. :) I added that to the calculator and uploaded it.

[Dagni]

Thanks, TheDragoon, for the test character. I tested only a few points, but enough that I think this is right.

It's very simple, 'WereBase' = 9 for Hunger (animation length of 10 frames). That's it. Only difference from normal. :) That's the case for either wolf or bear.

Ruvanal:

For the S3 sequence (biting attack animation) the lengths are WereBear=13, WereWolf=10.

How sure are you? :) A length of 10 for the Bear as well as the Wolf works perfectly with the points I tested.

- Dagni

[Ruvanal]

Ruvanal:
How sure are you? :)  A length of 10 for the Bear as well as the Wolf works perfectly with the points I tested.

- Dagni

Looking straight at the animation files. Single frame display or in slowed down cycle mode to see the action.

Wolf s3: 80 frames in 8 different facings. 80/8=10 per facing.

Bear s3: 208 frames in 16 different facings 208/16=13 per facing. On the bear, I am not sure that the last 8 facing are used since one of the files indicates that there should only be 8 facing used for this animations. Keep in mind that like the GH animations and BL animations that there may be a different speed factor used to keep these in close to the same speed.

Go to Phrozen Keep and get the CV5 program and the plugins for the DCC files and you can look for yourself.

Edit: I just did a non-timed test run of hunger with both wolf and bear using Fast, Very Fast and Slow weapons. While the speed difference with the Very Fast was hard to see, there was a difference in the speeds with Fast and a rater noticable difference in the speeds when using a Slow weapon. If it was working out that both animations were using the same value of 10 for the animation length, should not the speeds have appeared to be the same also?

[Ruvanal]

Edit: I just did a non-timed test run of hunger with both wolf and bear using Fast, Very Fast and Slow weapons.  While the speed difference with the Very Fast was hard to see, there was a difference in the speeds with Fast and a rater noticable difference in the speeds when using a Slow weapon.  If it was working out that both animations were using the same value of 10 for the animation length, should not the speeds have appeared to be the same also?

I made an error here (it is very late in a day for me). I had forgotten to take off the WereWolfs inherant IAS. When I took that off the speed were nearly identical, but I did see something that most would not even know about. When doing a short timing of the Bear with a slow weapon, I was seeing a "jerk" at the end of each animation cycle. Comparing that with a cycling of the main animation file it appears that for at least that case they are only using the first 10 frames of the 13 frame animation cycle to display the Bears' S3 animation. This would cause the quick jerk into an upright postion at the end of each bite attack as you would missing 3 frames of the bear finishing standing back up. Looks like bug to me.

[Dagni]

If it was working out that both animations were using the same value of 10 for the animation length, should not the speeds have appeared to be the same also?

No, the one (and only) difference is the 'prev frames' value of 10 for Bear, 9 for Wolf. That makes the bear faster.

Now, if for some reason the 'prev frames' value was different for Hunger with the WereBear than normal, then it equalling 13 that would have almost the exact same effect as the animation length = 10. I.e. 13/13 vs 10/10. The only difference is rounding. But the only value I found where the rounding differed tested out as the 10/10 one.

Keep in mind that like the GH animations and BL animations that there may be a different speed factor used to keep these in close to the same speed.

That's what I'm afraid of. Because without knowing just what - and where - that addition to the equation is, deriving a formula that won't be off in rare cases due to rounding errors is about impossible.

Could it be possible that the game treats it as if there are 10 frames, and only uses 10 even though 13 exist?

Thanks for the help

Ahh, while trying to get the stupid forum to post, I see the new post: so it IS possible that the game only uses 10 of the 13 frames!

- Dagni

[TheDragoon]

Ok, awesome work everyone! :lol: I tested about 20ish breakpoints with the formula and I'm convinced. Looks like we've got all the attacks covered! I've added that to the calculator and uploaded it.. it can be used here.

[Hammerman]

For the S3 animation (hunger), wolfs have anim speed = 240, and bears have 208. Both have animation lenght of 10, despite bears having 13 frames avalible in the DCC files.

Btw, a 'readable' dump of animdata.d2: http://www.geocities.com/hammermand2/files...es/animdata.txt

[TheDragoon]

The sound effects seem too long for the length of the attack, as well. I wonder if that is also part of the bug. :unsure:

[Ruvanal]

Hammerman, thanks for the information where to find that animation data (and the extracted file of it). I knew there had been some discussion in the past about this at the Phrozen Keep, but could find it recently, nor remembered any the keywords that would pull it up.

The sound effects seem too long for the length of the attack, as well. I wonder if that is also part of the bug. 

Since I have not even messed with the Hunger attack in about a year, I was not familiar with the sound to note that. It sounds like there was an error in using the wolf's animiation length over that of the bears. Yes there are puns in hear.

[Bolty]

I've added that to the calculator and uploaded it.. it can be used here.

Perfection is neared...

Great job everyone!

-Bolty

[TheDragoon]

Actually, I should be more specific. :) The sound effect weirdness is only for the Werebear. The reason I've noticed this is that I usually use the sound effects to help me with my animation frame number tests (I test the amount of time to swing 25 times.. the result, in seconds, is the number of rames) however I've found it MUCH easier to test Werebear Hunger attacks if I turn off the sounds since the sounds don't match up with the animations. The Werewolf Hunger attack sound effects are just fine. :P

[Dagni]

For the S3 animation (hunger), wolfs have anim speed = 240, and bears have 208. Both have animation lenght of 10, despite bears having 13 frames avalible in the DCC files.

Ok, first off, I just want to be sure I know which '256' to replace with that animation speed.

Now, normally 'Animation speed' is 256. But for wereform attack speeds it uses a differently calculated number... here's how. We begin with:

Animation speed = a1_frames*256 / ((weapon ias+wsm+100)*anim speed/100)

That 'anim speed' was being replaced with 256 in the formula before. So instead of:

Animation speed = a1_frames*256 / ((weapon ias+wsm+100)*256/100)

we use this for werebear's Hunger instead? Is that right?

Animation speed = a1_frames*256 / ((weapon ias+wsm+100)*208/100)

But, irregardless of where it goes, is it possible that this '208' isn't actually being used for some reason? Because I was getting test results that matched what the formula I was already using expected, and changing a '256' to '208' makes a large difference in the formula results. For Werewolf the difference is small enough that I may have simply not tested a point that was different, I would have to check. But for Werebear the difference would be obvious if that were the only change from what I was trying (and I was using S3 anim_length = 10).

- Dagni

[Hammerman]

Dagni,

Ok, first off, I just want to be sure I know which '256' to replace with that animation speed.

Now, normally 'Animation speed' is 256. But for wereform attack speeds it uses a differently calculated number... here's how. We begin with:

Animation speed = a1_frames*256 / ((weapon ias+wsm+100)*anim speed/100)

In this formula, you use the Druids human form A1 values, where anim speed is 256. The value I gave should go in the main speed formula instead of the latter * 256 part.

But, irregardless of where it goes, is it possible that this '208' isn't actually being used for some reason?

Doh... well ofcourse it should not be used, because when in shifted form that value is ignored. *slaps forhead*

Yea, ignore that 208. If you got the formula working, good job!

[TheDragoon]

In testing with the 1.10 Beta patch, Feral Rage seems to attack MUCH faster than it did before. It's often on par with Hunger. However, it's kind of hard to test since the attack length seems to be different whether you hit or miss the target.. I'm going to try and test some stuff experimentally for it, but can any code-readers out there look up and see if any of the animation length data has changed for the 1.10 beta patch? I don't really know where to find that sort of thing and it would probably be a big help in figuring out what is different. :)

Right now, the closest I've come to approximating what data I've taken is setting the wereform animation to 11 (as opposed to its original 13 and Hunger's 10) and subtracting an additional frame (so 2 overall) to get the actual animation length. It seems to match the data I've taken so far, but subtracting TWO frames off of the value doesn't sit very well with me. :)

[Nicodemus Phaulkon]

Is this why your Wereform Central hasn't been functional for a bit? I trundled over there to do some "comparison shopping" the other day and couldn't get the calc to spit anything out?

*tips helm*

[TheDragoon]

It's functional... just remember to click on one of the boxes to pick your Wereform or it'll give you errors (as it does kind of need to know whether you're dealing with a Werebear or a Werewolf ;) ).

[Nicodemus Phaulkon]

Radio buttons; even WITH HCI training, I always forget the damnable things. *sigh*

Thanks, TD.

*tips helm*