Kiss of the Spider - Page 3

Trazhenko · 08/21/06, 5:12 PM

Originally Posted by rochan

Originally Posted by Sarana

Ash @ 1540 RAP, 24.78% crit, vs. Level 63
Hit +7 DPS 3% hit
5 557.55 565.56 3% is better by 8.01 DPS
6 562.14 568.29 3% is better by 6.05 DPS
7 566.73 568.29 3% is better by 1.56 DPS
8 571.33 568.29 +7 is better by 3.04 DPS
9 574.09 568.29 +7 is better by 5.8 DPS

Is that +7 DPS a typo? (it should be +7 damage?)
Also how come the +7 has a diminishing effect? I would have thought it increases with %hit -- or is it just a glitch with the last entry?

I would imagine you could chalk it up to only needing 8.xx hit to never miss, so there is less potential damage to gain from adding the 9th +hit.

Sarana · 08/21/06, 10:10 PM

Hm, im getting different numbers now, i think i messed up somewhere... i'll fix it after my raid

Edit: I just checked it again, it seems correct to me. The default missrate vs a level 63 is ~8.6% so having 9% is ideal. The first line for the +3 scope DPS would be +8 (5+3), if that wasn't clear.

And yeah, +7 Damage. I meant to write +7 scope DPS, but forgot >_>

Hand · 10/01/06, 2:06 PM

I had a question reguarding this trinket, everyone seems to think its better to use it at the same time as blade flurry because attack speed multipliers are multiplicative, however wouldn't that mean they diminish as you stack them further, and it would be better to use them seperate? Feel free to correct my math here, I'm pretty sure I'm doing something wrong--

Assume a 1.0 speed weapon, and slice n dice is always up

1.0 / (1.3 * 1.2) = 0.641026 aspd with blade flurry snd
1.0 / (1.3 * 1.2 * 1.2) = 0.534188 aspd with all 3

0.641026 - 0.534188 = 0.106838 seconds off your swing timer from the KoTS

However, using it when blade flurry isn't up
1.0 / (1.3) = 0.769231 aspd with just snd
1.0 / (1.3 * 1.2) = 0.641026 aspd with snd and KoTS

0.769231 - 0.641026 = 0.128205 seconds off your swing timer from the KoTS

So for a rogue, would it not be better to use this seperate from blade flurry, as opposed to at the same time? Or am I missing something? As I said, I'm running under the assumption that my math is wrong, so I'm looking for someone to correct it. Thanks.

Eej · 10/01/06, 2:13 PM

Haste calculation works like this.

Weapon Speed / (1 + Haste Effects)

So:

1.0 / (1 + .30 + .20) = 0.67 attack speed with Blade Flurry and SnD
1.0 / (1 + .30 + .20 + .20) = 0.588 attack speed with Blade Flurry, SnD and KotS.

Hand · 10/01/06, 2:15 PM

I read that they were multiplicative, not additive o.o

So they are additive? Are you sure about this?

Zurai · 10/01/06, 3:10 PM

Eej is wrong. It's

Weapon Speed / (weapon haste 1 * weapon haste 2 * weapon haste 3 ...)

Which gives the correct, verifiable-in-game numbers of:

1.0/1.56 = 0.64 attack speed for BF + SnD
1.0/1.872 = 0.53 attack speed for BF + SnD + KotS

♦ Viper · 10/01/06, 3:19 PM

Originally Posted by Hand

I had a question reguarding this trinket, everyone seems to think its better to use it at the same time as blade flurry because attack speed multipliers are multiplicative, however wouldn't that mean they diminish as you stack them further, and it would be better to use them seperate? Feel free to correct my math here, I'm pretty sure I'm doing something wrong--

Assume a 1.0 speed weapon, and slice n dice is always up

1.0 / (1.3 * 1.2) = 0.641026 aspd with blade flurry snd
1.0 / (1.3 * 1.2 * 1.2) = 0.534188 aspd with all 3

0.641026 - 0.534188 = 0.106838 seconds off your swing timer from the KoTS

However, using it when blade flurry isn't up
1.0 / (1.3) = 0.769231 aspd with just snd
1.0 / (1.3 * 1.2) = 0.641026 aspd with snd and KoTS

0.769231 - 0.641026 = 0.128205 seconds off your swing timer from the KoTS

So for a rogue, would it not be better to use this seperate from blade flurry, as opposed to at the same time? Or am I missing something? As I said, I'm running under the assumption that my math is wrong, so I'm looking for someone to correct it. Thanks.

You can model the total number of attacks per unit time to get a better picture of how it works. 1.3 weapon haste on a 1.0 speed weapon = 1.0/1.3 aspd = 1.3 attacks per second, as intuition would tell you.

So suppose the buffs all have equal durations and you want to model a period of time that is twice the duration.

Using BF + SnD, then KotS after: (1.0)(1.3)(1.2) + (1.0)(1.2) = 2.76 attacks
Using all 3 at once, then none after: (1.0)(1.3)(1.2)(1.2) + (1.0) = 2.872 attacks

So yes, using all of them at once is still better.

And to explain your own results: Yes there are diminishing returns on your swing timer, but the swing timer is an inverse function of the actual damage you do. Going from 1.0 to 0.8 on the swing timer changes your attack rate from 1.0 to 1.25, a 25% increase for a 0.2 reduction in swing time. But going from 0.5 to 0.4 on your swing timer changes your attack rate from 2.0 to 2.5-- also a 25% increase, but for only a 0.1 reduction in swing time.

Hand · 10/01/06, 3:24 PM

Ah, theres the answer I was looking for. :)

Thanks a bunch Viper.

Decker · 10/01/06, 3:34 PM

Just for argument sake (based on smiting XBow)

Quiver: 15%
Rapid Fire: 40%
KOTS: 20%
IAOTH: 30%
Beserking: 10%-30%

Normal Hunters: (all but Zerking) 1.35 weapon speed
Trolls: 1.23 to 1.04 weapon speed

These are actually practical buffs too ;)

Eej · 10/01/06, 3:38 PM

Ah here we go, looks like I screwed up.

Originally Posted by Das 1.12 PTR notes

- Haste and Slow effects
Previously Haste and Slow effects worked inconsistently, with spells
working differently from weapons, and hastes and slows not acting as
inverses of each other. We have revised the system so that all haste
and slow effects work the same way, and haste and slow percentages
perfectly cancel each other out (30% haste and 30% slow combine to
no change). As a result, we had to change the tooltip numbers on all
spell haste effects, and on all melee and range slow effects. The
numbers in the tooltips are different, but the game functionality
is unchanged (other than slight rounding errors). Those tooltips
that changed will now display larger numbers than they used to
display. Conceptually, haste values indicate how much more of that
activity you can perform in a given time. 30% melee haste means 30%
more swings in a given time. Slow values indicate how much longer an
activity takes to complete. 30% slow means an action takes 30%
longer to finish. For those interested, the new haste and slow
effects work mathematically in the following way. Each effect
separately multiplies the attack timer of your weapon or casting
time of your spell. They are multiplied in sequence (order doesn't
matter) and the final result is the attack timer or casting time.
The multipliers use the following equations:

Haste effects: multiplier=1/(1+percent/100)
Slow effects: multiplier=1*(1+percent/100)

As an example, a 50% haste generates a multiplier of 0.666; a 50%
slow generates a multiplier of 1.5. If they are multiplied together
the result is 1. A 50% haste on a 3.0 speed weapon makes it 2.0
speed, resulting in you getting 30 swings per minute instead of 20
swings per minute: 50% more swings. A 50% slow on a 3.0 casting time
spell generates a casting time of 4.5. It took you 50% longer to
finish casting the spell.

That gives me 1.92 attack speed with quiver + rapid fire + berserking, which is close enough to 1.90 for me.

Boevis · 10/01/06, 3:44 PM

Well, any good rogue will have SnD up all the time, so it would look more like:
BF+SnD then KotS+SnD (1.0)(1.3)(1.2) + (1.0)(1.3)(1.2) = 3.12 attacks
BF+SnD+KotS then SnD (1.0)(1.3)(1.2)(1.2) + (1.0)(1.3) = 3.172 attacks

So it's only a .052 difference instead of a .112, but the end result is still stacking is better. KotS and BF have the same cooldown anyway, so it works very well and reliably.

I know it's just theorycrafting, but I wouldn't be saving my cooldowns waiting for a IAotH proc ;p