Combat Mechanics? Attack Speed, Recovery

[Warden]

the only thing i could think of that might cause this is fraps or virtual dub.

 

do you have the "lock frame rate while recording" option on ?

because if your game runs at say 45 and you record at 30 it will be highly inaccurate

 

so either lock it in fraps or in drivers

[kmbogd]

I think that someone mentioned that all modifiers for damage are first added together and then multiplied by the base damage. This is not true for weapon quality modifier. The modifier of the weapon quality is applied to the damage of the basic weapon and this becomes the base damage. Now you use the everything additive then multiplicative rule. This is based on the damage information provided in the character information sheet (hopefully the in battle implementation is not different).

 

ps: did you manage to test the findings of my previous post?

Edited April 3, 2015 by kmbogd

[Baki]

 

the only thing i could think of that might cause this is fraps or virtual dub.

 

do you have the "lock frame rate while recording" option on ?

because if your game runs at say 45 and you record at 30 it will be highly inaccurate

 

so either lock it in fraps or in drivers

 

yep

EDIT: wait..  i had the videop capture settings on 30 fps but the framerate was not locked.. okay, all again :D

@kmbogd: nope, ill try to figure out the ranged stuff first

Edited April 3, 2015 by Baki

[hugin7]

I made post about Penetrating shot, but hardly anyone saw it, so:

 

 

Penetrating Shot talent work for spells.

Tested:

 

 

Cipher:

 

Affected (DR reduced by 5):

 

1) Mind Blades

 

2) Mind Lance

 

 

 

Not affected (ones that do not target enemy)

 

1) Soul shock

 

2) Antipathetic Field

 

3) Ectopsychic Echo

 

4) Amplified Thrust

 

 

 

Wizard

 

 

 

Affected (DR reduced by 5):

 

1) Thrust of Tattered Veils

 

2) Necrotic Lance

 

3) Crackling Bolt

 

4) Minoletta's Minor Missiles

 

5) Minoletta's Bounding Missiles

 

6) Minoletta's Concussive Missiles

 

7) Rolling Flame

 

 

 

Not affected

 

1) Fan of Flames

 

2) Ghost Blades

 

3) Kalakoth's Sunless Grasp

 

 

 

 

 

So... it seems Penetrating Shot talent affect powers/spells that

 

- single target (also ones that jump from enemy to enemy)

 

- straight line aoe

 

 

 

Which makes Penetrating Shot talent interesting choice for casters.

 

 

 

 

 

ps:

 

Also Vulnerable Attack talent affect Wizard's touch spells:

 

1) Kalakoth's Sunless Grasp

 

2) Jolting Touch

 

 

 

pps:

 

I havnt tested Druid, and Cleric dont have direct damaging spells (one that line aoe seems not working)

 

ppps:

 

Could anyone frapstest how Pen Shot affect casting time?

 

I tested a fairly typical case for Penetrating Shot via Fraps/MPH-HC, 30 frames/sec:

 

Cipher, 20 dex, Gunner, and reload chant.  I used a blunderbuss.

 

In this scenario, there's a 9.7% overall rate of fire gain for avoiding Penetrating Shot.  If weapon damage > 50 (typical with accuracy bonus > 10, arbalest, damage mod > 1.5, and DR debuff > 5), penetrating shot leads to a DPS loss for most combat scenarios, not to mention focus loss for Ciphers due to reduced rate of fire.  For arbalests fired against DR = 20 targets, there is about a 2.5% DPS increase with Penetrating Shot.

 

A combination of poor buffs/debuffs, low damage weapons, and high DR targets can make penetrating shot viable as a situational DPS booster, but it seems that such cases are uncommon, so our talent points are likely better spent elsewhere.  Ironically, blunderbuss-wielding Ciphers, who tend to have the most to gain from Penetrating Shot in terms of weapon damage, also have the most to lose because of focus issues.

 

Things got more interesting when I tested casting spells - there is no casting speed loss with Penetrating Shot.  However, since only a few spells benefit from DR, the talent remains of dubious value.  Ciphers still have the issue of focus loss, while other casters are almost always better off casting CC, buffs, and debuffs to flip attack resolution tables in your party's favor and render enemies useless.

 

Of course, your mileage may vary depending on the situation/strategy.

[ErlKing]

Thx for testing.

 

Well, this talent is toggleable, so...  getting -5 DR for Mind Blades, Mind Lance is pretty neat. Also, you sure that blunderbuss suffer from Pen Shot talent? 6*5=30 extra damage, not such bad bonus.

Edited April 3, 2015 by ErlKing

[AlphaMagnum]

I think that someone mentioned that all modifiers for damage are first added together and then multiplied by the base damage. This is not true for weapon quality modifier. The modifier of the weapon quality is applied to the damage of the basic weapon and this becomes the base damage. Now you use the everything additive then multiplicative rule. This is based on the damage information provided in the character information sheet (hopefully the in battle implementation is not different).

 

ps: did you manage to test the findings of my previous post?

Unfortunately my tests contradict what you said. It's true that weapon tooltips suggest that you're right, but the damage range displayed in the character sheet proves that the weapon tooltip is just misleading. Here's an example:

 

2H estoc - 14-20 base damage

Superb is +45% so 20.3 - 29, game displays 20-29.

Say your Might is 20, so 30% extra damage.

 

Case 1 - Superb creates a "new" base damage

Damage should be (20.3) * 1.3 to 29 * 1.3

26.39 - 37.7, rounding to 26-38

 

Case 2 - Everything is additive

Damage should be (14) * (1 + .45 + .3) to (20) * (1 + .45 + .3)

14 * 1.75 to 20 * 1.75

24.5 - 35, rounding to 25-35

 

If you look in game, the character sheet displays 25-35, showing that Case 2 is how the game handles these calculations.

Edited April 3, 2015 by AlphaMagnum

[Baki]

okay, this time with locked framerate

 

Base

AT 52

REC 74

REL  98

 

Only Chant (Rec x1.2, REL x2)

AT 52

REC 60 -> 74*(1-0.2)=59.2

REL 51 = 98*(1-0.5)=49

 

Only Armor (Rec -50%)

AT 52  

Rec 112 -> 74*(1+0.5)=111

Rel 98

 

Only Dex (AT, REC, REL 24%)

AT 44  -> 52*(1-0.24)=39.52  <- wrong

REC 60 <- 74*(1-0.24)=56.24 <- wrong

REL 81 <- 98*(1-0.24)=74.48 <- wrong

 

Only Pen Shot (Rec -20%)

AT 52

REC 89 -> 74*(1+0.2)=88.8

REL 98

 

ALL

AT 44  -> 52*(1-0.24)=39.52  <- wrong

REC 90  -> 74*(1-0.24+0.2+0.5-0.2)=93.24  <- wrong

REL 43 -> 98*(1-0.5-0.24)=25.48  <- wrong, maybe multiplicative stacking?

 

Any suggestions?

[hugin7]

Thx for testing.

 

Well, this talent is toggleable, so...  getting -5 DR for Mind Blades, Mind Lance is pretty neat. Also, you sure that blunderbuss suffer from Pen Shot talent? 6*5=30 extra damage, not such bad bonus.

 

Damage-wise, Penetrating Shot is tailor-made for blunderbusses.  Unfortunately, most characters besides Ciphers get much better results with arbalests in almost all situations, and Ciphers mainly use it for focus gain.  Lower attack rate means less focus, so there is a trade-off for Ciphers.

[Zaruthustran]

How does bow (or wand, or other fast ranged weapon with no reload) compare to blunderbuss, for focus gain? Blunderbuss is 6 shots at once, but it seems that during the long reload time you could get off several bow shots.

[kmbogd]

 

I think that someone mentioned that all modifiers for damage are first added together and then multiplied by the base damage. This is not true for weapon quality modifier. The modifier of the weapon quality is applied to the damage of the basic weapon and this becomes the base damage. Now you use the everything additive then multiplicative rule. This is based on the damage information provided in the character information sheet (hopefully the in battle implementation is not different).

 

ps: did you manage to test the findings of my previous post?

Unfortunately my tests contradict what you said. It's true that weapon tooltips suggest that you're right, but the damage range displayed in the character sheet proves that the weapon tooltip is just misleading.

 

If you look in game, the character sheet displays 25-35, showing that Case 2 is how the game handles these calculations.

 

 

Well I actually looked at the character sheet and derived what I said. By the way, I believe that the tool tip is not actually rounding but it's displaying only the integer value. If it was rounding to the nearest integer than there are some serious bugs in there according to what I've seen. So let me provide you with the full example of how I arrived to my conclusion.

 

From the picture below you can see that my character has 21 Might which equates to 33% damage.

 

 

I also have a fine estoc equiped which has a 1.15 dmg increase with respect to the standard 14-20 dmg range:

 

 

As you can see above both the character sheet and the equipped weapon sheet display a 21 - 30 range. Let's apply the 2 rules:

 

1) weapon modifier applied separately to provide base dmg:

 

[14-20]*1.15=[16.1 - 23]; this actually corresponds to the base dmg of the equipped weapon sheet.

 

Now I apply my 33% might modfier and I get [16.1 - 23]* (1 + 0.33)=[21.413 - 30.59]

 

The integer parts correspond exactly to the damage range displayed in either of the sheets.

 

 

2) weapon modifier applied additively like the other modifiers

[14-20]*(1+0.15+0.33)= [20.72 - 29.6]

 

The integer parts don't correspond to the damage range displayed.

 

 

The only point of arguing is whether the tooltip rounds to the nearest integer or displays only the integer part. I am almost sure that it displays the integer part but if you want I could start working on doing tests on that. Btw did you double check the numbers in your example, do you have a screenshot of that?

[AlphaMagnum]

Don't have screenies (on mobile) but I tested with multiple sources of bonus damage.

 

Your example doesn't give a huge differential. Additive means total 1.48x base, whole multiplicative is 1.53. Not a huge difference.

 

I tested with 20 Might (30%), Superb weapon (45%), and Reckless Assault (20%).

Additive - 1.95x base

Superb modified base damage - (1.45 * 1.5) = 2.175x base damage

 

This difference of 1.95 vs 2.175 produces much more substantial differences in the final results, which you can see in the character sheet. There's no way you can confuse 39 and 43, regardless of how your system rounds, assuming that it is competently performing the calculations.

[AlphaMagnum]

And I did indeed test my theory in game as outlined in the post above. Just didn't take screenshots.

[Myrten]

okay, this time with locked framerate

 

Base

AT 52

REC 74

REL  98

 

Only Chant (Rec x1.2, REL x2)

AT 52

REC 60 -> 74*(1-0.2)=59.2

REL 51 = 98*(1-0.5)=49

 

Only Armor (Rec -50%)

AT 52  

Rec 112 -> 74*(1+0.5)=111

Rel 98

 

Only Dex (AT, REC, REL 24%)

AT 44  -> 52*(1-0.24)=39.52  <- wrong

REC 60 <- 74*(1-0.24)=56.24 <- wrong

REL 81 <- 98*(1-0.24)=74.48 <- wrong

 

Only Pen Shot (Rec -20%)

AT 52

REC 89 -> 74*(1+0.2)=88.8

REL 98

 

ALL

AT 44  -> 52*(1-0.24)=39.52  <- wrong

REC 90  -> 74*(1-0.24+0.2+0.5-0.2)=93.24  <- wrong

REL 43 -> 98*(1-0.5-0.24)=25.48  <- wrong, maybe multiplicative stacking?

 

Any suggestions?

It's clearly dex bonus which breaks everything - actual decrease is like 18% instead of 24% - can you do some tests with different dex values? This way maybe we could figure out some formula

[kmbogd]

The information sheets round to the nearest integer, they don't display only the integer value as I thought they would. So AlphaMagnum is right, the weapon quality modifier is also applied like the rest.

Edited April 3, 2015 by kmbogd

[MaximKat]

Looks like the Dex bonus is applied as AttackTime/(1+Bonus), instead of AttackTime*(1-Bonus) as the chant does. E.g., 74/1.24=59.67 - matches observed.

 

For the ALL case, REC = 74*(1+0.2+0.5-0.2)/(1+0.24)=89.5 - also works.

 

EDIT: fixed a typo in the numbers.

Edited April 3, 2015 by MaximKat

[MaximKat]

Also, considering the small (couple of frames) variation between measurements, reported in this thread, does it seem very likely that many weapons in the spreadsheet actually have the same timings, instead of a bunch of different values in the 50-54 range? Could someone verify it and clean the data up? E.g., all slow ranged weapons are actually 50/75.

[Baki]

Looks like the Dex bonus is applied as AttackTime/(1+Bonus), instead of AttackTime*(1-Bonus) as the chant does. E.g., 74/1.24=57.67 - matches observed.

 

For the ALL case, REC = 74*(1+0.2+0.5-0.2)/(1+0.24)=89.5 - also works.

57.65 is too far from 60 imo.

 

~2.35/14=16.8% deviation

 

Edit: oh, 74/1.24 is actually 59.68

 

GJ MaximCat!

 

also, you are welcome to do the data cleanup yourself

 

fraps and virtualdub are everything you need.

 

Another edit: i added MaximKat's attack speed findings in the sheet. Also edited Melee Fast and Arbalest Dmg (Patch)

Edited April 3, 2015 by Baki

[Baki]

Can we really be so sure that MaximKat's formula is correct?

 

-Does not work for Attack Speed: 52/1.24=42 and not 44 as expected

-Does not work for Reload Speed: 98/1.24= 79 and not 81 as expected

[MaximKat]

Oops, yeah, that was a typo. Was supposed to be 59.67.

 

I'm more of a theory person Would be happy to help analyze the data, but recording is not my thing.

[Baki]

next time please come up with a theory that works for all cases before declaring it valid

[MaximKat]

Can we really be so sure that MaximKat's formula is correct?

 

-Does not work for Attack Speed: 52/1.24=42 and not 44 as expected

-Does not work for Reload Speed: 98/1.24= 79 and not 81 as expected

It's probably just precision issues in the data. For example, 98*0.5 is 49, not 51. I don't think errors of 1-2 frames are meaningful.

[Baki]

im not so sure about that. but its the closest we have so far.

 

this is so frustrating. i will never understand why devs do not reveal this stuff. 

Edited April 3, 2015 by Baki

[AlphaMagnum]

Slightly off-topic: Anyone in-game had a chance to see how 1.03 modified fast 1h weapons and sabres?  I can't get in right now but I'm curious as to whether the improvement was substantial enough to shake up game balance.

[Daemonjax]

Here's data I gathered for 1h dagger (not dual-wielding) using a dex 10 naked character:

 

 

 

50 : Game is paused

51 : Game is NOT paused

-------------------------------

52 : Actor animations start playing 

-------------------------------

56 : Attack animation_A begins ( 56 -  51 =  5 frame attack delay)

73 : Idle animation begins ( 73 -  56 = 17 frame attack animation_A length)

78 : Recovery bar is full       ( 78 -  73 =  5 frame recovery delay)

110: Recovery bar is empty (110 -  78 = 32 frame recovery length)

------------------------------- (110 -  51 = 59 frames total

 

115: Attack animation_B begins (115 - 110 =  5 frame attack delay)

133: Idle animation begins (133 - 115 = 18 frame attack animation_B length)

137: Recovery bar is full (137 - 133 =  4 frame recovery delay)

168: Recovery bar is empty (168 - 137 = 31 frame recovery length)

------------------------------- (168 - 110 = 58 frames total

 

173: Attack animation_A begins (173 - 168 =  5 frame attack delay)

190: Idle animation begins (190 - 173 = 17 frame attack animation_A length)

196: Recovery bar is full (196 - 190 =  6 frame recovery delay)

227: Recovery bar is empty (227 - 196 = 31 frame recovery length)

------------------------------- (227 - 168 = 59 frames total

 

232: Attack animation_A begins (232 - 227 =  5 frame attack delay)

249: Idle animation begins (249 - 232 = 17 frame attack animation_A length)

255: Recovery bar is full (255 - 249 =  6 frame recovery delay)

286: Recovery bar is empty (286 - 255 = 31 frame recovery length)

------------------------------- (286 - 227 = 59 frames total

 

291: Attack animation_A begins (291 - 286 =  5 frame attack delay)

308: Idle animation begins (308 - 291 = 17 frame attack animation_A length)

314: Recovery bar is full (314 - 308 =  6 frame recovery delay)

345: Recovery bar is empty (345 - 314 = 31 frame recovery length)

------------------------------- (345 - 286 = 59 frames total

 

350: Attack animation_B begins (350 - 345 =  5 frame attack delay)

368: Idle animation begins (368 - 350 = 18 frame attack animation_B length)

372: Recovery bar is full (372 - 368 =  4 frame recovery delay)

403: Recovery bar is empty (403 - 372 = 31 frame recovery length)

------------------------------- (403 - 345 = 58 frames total

 

408: Attack animation_A begins (408 - 403 =  5 frame attack delay)

425: Idle animation begins (425 - 408 = 17 frame attack animation_A length)

431: Recovery bar is full (431 - 425 =  6 frame recovery delay)

462: Recovery bar is empty (463 - 431 = 32 frame recovery length)

------------------------------- (462 - 403 = 59 frames total

 

 

 

 

462 - 51 =  411 Total frames

411 /  7 = 58.7 Average frames per attack

 

However, I would use (59 + 58) / 2 = 58.5 average frames per attack since there's probably a 50/50 chance of using either of the two attack animations.

Edited April 3, 2015 by Daemonjax

[Myrten]

Please be careful when you edit the spreadsheet, someone has deleted crit chance for zero accuracy mod